Merge branch 'mai/kahn/phase-6-sliceB' (Phase 6 Slice B: mBrian-backed read path behind PROJAX_BACKEND switch, parity-green)

feat(adapter): Phase 6 Slice B — mBrian-backed read path live
Per t-projax-6-sliceB-readpath. mBrian migration (m/mBrian#73) is live on msupabase with 65 nodes + 78 child_of + 81 projax-* edges. This commit makes the projax read path source from there behind an env switch. CLIENT ARCH: direct pgxpool against mbrian.* schema (same SUPABASE_DATABASE_URL the projax binary already uses for projax.*) — matches flexsiebels/head's cross-coupling pattern. No MCP token plumbing. CONTRACT (all three honoured) - External links are SELF-EDGES (source=target=item, rel='projax-*', payload in edges.metadata). linkFromEdge reads the node's outbound projax-* edges; ref_id derived per ref_type from metadata (caldav url, gitea owner/repo, mai-project mai_project_id). - Slugs finalised: 'work'/'dania' resolve to mBrian's canonical nodes; projax-side squatters (renamed-aside, not deleted) are documented in the parity test as legacy-only and skipped from field comparison. - created_at/updated_at NOT preserved — ItemsCreatedInRange orders off metadata.projax.start_time when present, fall back to mBrian created_at. Aggregator surfaces (timeline / dashboard) read off caldav DTSTART + gitea updated_at, so they're unaffected. NEW FILES - store/mbrian.go: MBrianReader concrete impl. Bulk-loads projax- managed nodes + child_of edges in one pair of queries per call, builds a graphContext in memory, derives Paths via ancestor walk (depth-capped at 64 like projax's trigger). Implements every ItemReader method. - store/mbrian_parity_test.go: 5 parity tests against the live db — ListAll field equality (skipping the renamed squatter slugs), spot-check resolves, caldav-list link round-trip, gitea-repo link round-trip, AllTags union, NotFound consistency. All 5 GREEN. - cmd/projax-remap-views/main.go: one-shot tool to rewrite projax.views.filter_json.project_id from old projax uuids to new mBrian uuids using the audit map mBrian dropped (head will relay the path). Dry-run default; --apply commits. Idempotent. - docs/plans/slice-b-views-projectid-gap.md: surfaces the gap + the remediation path. Must run remap BEFORE slice E drops projax.items. CHANGES - store/adapter.go: kept the ItemReader interface + *Store assertion; removed the prep stub (replaced by mbrian.go). - web/server.go: Server.Items store.ItemReader field. web.New defaults Items to the concrete *Store (legacy path). main.go overrides to MBrianReader when PROJAX_BACKEND=mbrian. - All read-path call sites in web/ swapped from s.Store.<readMethod>( to s.Items.<readMethod>( for the 15 ItemReader methods. MCP tools unchanged (separate scope; can pivot in a follow-up). Writes still flow through s.Store. - cmd/projax/main.go: PROJAX_BACKEND env switch with "store" (default) and "mbrian" values. Logs the choice at startup. Unknown value refuses to start. SMOKE - go build ./... green; go vet green. - go test ./store/ -count=1 — all parity tests pass against live data. - Local server boot with PROJAX_BACKEND=mbrian — backs binding logs "backend=mbrian (read path via store.MBrianReader)" and serves /views/tree (auth wall protects deeper smoke; parity tests cover that surface). PRE-EXISTING failure NOT addressed in this commit: 3 timeline_filter tests in web/ already failed on main (legacy /timeline URL hits the Phase 5j 301 redirect to /views/timeline). No diff vs main in those test files; out of scope for slice B. OUT OF SCOPE FOR SLICE B (deferred): - MCP read tools migration to ItemReader (separate diff, low risk). - Aggregator's LinkLister wired to ItemReader (currently consumes *Store directly through Server.Aggregator()). - views.filter_json.project_id remap RUN — tool ships here, run waits on the head's relay of the audit-map path. - Slice C write-path. Slice D mai-bridge worker. Slice E drop.
2026-05-31 22:22:40 +02:00 · 2026-05-31 22:20:38 +02:00 · 2026-05-29 15:18:15 +02:00 · 2026-05-29 15:17:24 +02:00
18 changed files with 1837 additions and 45 deletions
--- a/cmd/projax-remap-views/main.go
+++ b/cmd/projax-remap-views/main.go
@@ -0,0 +1,139 @@
+// projax-remap-views rewrites projax.views.filter_json.project_id from
+// the OLD projax.items uuid to the new mBrian node uuid using the audit
+// map mBrian dropped after the migration. Dry-run by default; pass
+// --apply to commit.
+//
+// Phase 6 Slice B gap remediation — see
+// docs/plans/slice-b-views-projectid-gap.md for the surrounding context.
+//
+// Usage:
+//
+//	projax-remap-views --map /path/to/uuid-map.json            # dry-run
+//	projax-remap-views --map /path/to/uuid-map.json --apply    # commit
+//
+// Map shape: {"<old-uuid>": "<new-uuid>", ...}
+//
+// Env: PROJAX_DB_URL or SUPABASE_DATABASE_URL — direct postgres URL.
+package main
+
+import (
+	"context"
+	"encoding/json"
+	"flag"
+	"fmt"
+	"os"
+	"time"
+
+	"github.com/jackc/pgx/v5/pgxpool"
+)
+
+func main() {
+	mapPath := flag.String("map", "", "JSON file with {old-uuid: new-uuid} map")
+	apply := flag.Bool("apply", false, "commit changes (default: dry-run)")
+	flag.Parse()
+
+	if *mapPath == "" {
+		die("--map is required")
+	}
+	mp, err := loadMap(*mapPath)
+	if err != nil {
+		die("load map: %v", err)
+	}
+	fmt.Fprintf(os.Stderr, "loaded %d uuid mappings\n", len(mp))
+
+	dbURL := os.Getenv("PROJAX_DB_URL")
+	if dbURL == "" {
+		dbURL = os.Getenv("SUPABASE_DATABASE_URL")
+	}
+	if dbURL == "" {
+		die("set PROJAX_DB_URL or SUPABASE_DATABASE_URL")
+	}
+
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+	pool, err := pgxpool.New(ctx, dbURL)
+	if err != nil {
+		die("pool: %v", err)
+	}
+	defer pool.Close()
+
+	rows, err := pool.Query(ctx,
+		`SELECT slug, filter_json::text FROM projax.views WHERE filter_json ? 'project_id'`)
+	if err != nil {
+		die("query: %v", err)
+	}
+	defer rows.Close()
+	type view struct {
+		slug    string
+		filter  map[string]any
+	}
+	var pending []view
+	for rows.Next() {
+		var slug, raw string
+		if err := rows.Scan(&slug, &raw); err != nil {
+			die("scan: %v", err)
+		}
+		var fj map[string]any
+		if err := json.Unmarshal([]byte(raw), &fj); err != nil {
+			fmt.Fprintf(os.Stderr, "skip %s: invalid JSON: %v\n", slug, err)
+			continue
+		}
+		oldID, _ := fj["project_id"].(string)
+		newID, ok := mp[oldID]
+		if !ok {
+			fmt.Fprintf(os.Stderr, "skip %s: project_id %q not in map (possibly already remapped)\n", slug, oldID)
+			continue
+		}
+		fj["project_id"] = newID
+		pending = append(pending, view{slug: slug, filter: fj})
+		fmt.Fprintf(os.Stderr, "  %s: %s → %s\n", slug, oldID, newID)
+	}
+
+	if len(pending) == 0 {
+		fmt.Fprintln(os.Stderr, "nothing to remap")
+		return
+	}
+
+	if !*apply {
+		fmt.Fprintf(os.Stderr, "DRY RUN — %d view(s) would be remapped; pass --apply to commit\n", len(pending))
+		return
+	}
+
+	tx, err := pool.Begin(ctx)
+	if err != nil {
+		die("begin: %v", err)
+	}
+	defer tx.Rollback(ctx)
+	for _, v := range pending {
+		payload, err := json.Marshal(v.filter)
+		if err != nil {
+			die("marshal %s: %v", v.slug, err)
+		}
+		if _, err := tx.Exec(ctx,
+			`UPDATE projax.views SET filter_json = $1::jsonb WHERE slug = $2`,
+			string(payload), v.slug); err != nil {
+			die("update %s: %v", v.slug, err)
+		}
+	}
+	if err := tx.Commit(ctx); err != nil {
+		die("commit: %v", err)
+	}
+	fmt.Fprintf(os.Stderr, "committed %d remap(s)\n", len(pending))
+}
+
+func loadMap(path string) (map[string]string, error) {
+	b, err := os.ReadFile(path)
+	if err != nil {
+		return nil, err
+	}
+	out := map[string]string{}
+	if err := json.Unmarshal(b, &out); err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
+func die(format string, args ...any) {
+	fmt.Fprintf(os.Stderr, format+"\n", args...)
+	os.Exit(1)
+}
--- a/cmd/projax/main.go
+++ b/cmd/projax/main.go
@@ -7,6 +7,7 @@ import (
 	"net/http"
 	"os"
 	"os/signal"
+	"strings"
 	"syscall"
 	"time"

@@ -65,12 +66,31 @@ func main() {
 		logger.Info("migrations applied")
 	}

-	srv, err := web.New(store.New(pool), logger)
+	st := store.New(pool)
+	srv, err := web.New(st, logger)
 	if err != nil {
 		logger.Error("server init", "err", err)
 		os.Exit(1)
 	}
 	srv.Version = gitCommit
+
+	// Phase 6 Slice B — backend selector. PROJAX_BACKEND=mbrian flips the
+	// read path to the mBrian-backed adapter; default keeps the legacy
+	// pgx-against-projax.items path so production rollback is one env
+	// flip. Writes still flow through srv.Store either way.
+	backend := strings.ToLower(strings.TrimSpace(os.Getenv("PROJAX_BACKEND")))
+	switch backend {
+	case "mbrian":
+		srv.Items = store.NewMBrianReader(pool)
+		logger.Info("backend=mbrian (read path via store.MBrianReader)")
+	case "", "store":
+		// Default — srv.Items is the *Store from web.New.
+		logger.Info("backend=store (read path via legacy *store.Store)")
+	default:
+		logger.Error("unknown PROJAX_BACKEND value", "value", backend)
+		os.Exit(1)
+	}
+
 	logger.Info("startup", "version", gitCommit)

 	if supaURL := os.Getenv("SUPABASE_URL"); supaURL != "" {
--- a/docs/plans/slice-b-adapter-contract.md
+++ b/docs/plans/slice-b-adapter-contract.md
@@ -0,0 +1,228 @@
+# Phase 6 Slice B — read-path adapter contract
+
+**Status**: prep work (this doc). No implementation.
+**Branch**: `mai/kahn/phase-6-sliceB-prep`.
+**Author**: kahn (coder, prep mode), 2026-05-29.
+**Parent plan**: `docs/plans/mbrian-backend-migration.md` (on `main`).
+**Scope boundary**: contract + compile-checking skeleton only. The mBrian-backed implementation waits on m/mBrian#73 landing the migration + handing over the uuid map.
+
+---
+
+## §1 — Consumer inventory
+
+Every read-path call site against `*store.Store` and the projax-shaped `Item` / `ItemLink` types. The interface (§2) is the union of these.
+
+### §1.1 — `*store.Store` read methods (source: `store/store.go`)
+
+| method | signature | semantics |
+|---|---|---|
+| `ListAll` | `(ctx) ([]*Item, error)` | every live item, ordered by `paths NULLS FIRST, slug` |
+| `GetByID` | `(ctx, id) (*Item, error)` | single item by uuid |
+| `GetByPath` | `(ctx, path) (*Item, error)` | resolve `dev.paliad` style path to leaf item |
+| `GetByPathOrSlug` | `(ctx, key) (*Item, error)` | path first, fall back to bare slug |
+| `Roots` | `(ctx) ([]*Item, error)` | items with `cardinality(parent_ids) = 0` |
+| `MaiOrphans` | `(ctx) ([]*Item, error)` | mai-managed root items needing classify |
+| `ListByFilters` | `(ctx, SearchFilters) ([]*Item, error)` | structured search (status / mgmt / has-link / paths-prefix) |
+| `Search` | `(ctx, q, limit) ([]*Item, error)` | trigram + FTS title/content/aliases |
+| `AllTags` | `(ctx) ([]string, error)` | union of every item's tags |
+| `LinksByType` | `(ctx, itemID, refType) ([]*ItemLink, error)` | one item's links of a given `ref_type` (empty = all) |
+| `LinksByRefType` | `(ctx, refType) ([]*ItemLink, error)` | every link of a given ref_type across items |
+| `DatedLinks` | `(ctx, itemID) ([]*ItemLink, error)` | one item's links anchored to a date (PER artifacts) |
+| `DatedLinksRange` | `(ctx, from, to) ([]*ItemLinkWithItem, error)` | dated links within window, joined with their item |
+| `RecentDocuments` | `(ctx, since, limit) ([]*ItemLinkWithItem, error)` | recent dated docs, joined with their item |
+| `ItemsCreatedInRange` | `(ctx, from, to) ([]*Item, error)` | items created within window |
+
+### §1.2 — Consumer call sites (by file)
+
+Each row = one read-path call site. Direct Pool access (admin.go counts, bulk.go filter-tx, links.go event-date update) is flagged separately at the bottom — those rework targets are out of slice B's read-path scope.
+
+| consumer | method | use case |
+|---|---|---|
+| `web/server.go handleTree` | `ListAll`, `AllTags`, `linkKindsByItem` (LinksByRefType ×N) | render /views/tree with chip-counted forest |
+| `web/server.go handleDetail` | `GetByPath` ×2 (PER fallback), `LinksByType` (caldav), `DatedLinks` | render /i/{path} detail page |
+| `web/server.go parentOptions` | `ListAll` | populate parent <select> on /new + /reparent |
+| `web/server.go handleClassify` | `MaiOrphans`, `parentOptions` | render /admin/classify |
+| `web/dashboard.go handleDashboard` | `ListAll`, `LinksByRefType` (caldav), `LinksByType` (gitea) ×N, `RecentDocuments` | Tiles + tasks + events + docs cards |
+| `web/calendar.go handleCalendar` | `ListAll` | month grid scope |
+| `web/timeline.go handleTimeline` + `buildTimeline` | `ListAll`, `linkKindsByItem` | chronological spine |
+| `web/graph.go handleGraph` | `ListAll`, `AllTags` | DAG SVG render |
+| `web/bulk.go handleBulk` | `ListAll`, `AllTags`, `GetByID` | /admin/bulk filtered checklist |
+| `web/caldav.go` (admin + create/unlink) | `ListAll`, `LinksByRefType`, `LinksByType`, `GetByPath` | /admin/caldav surface |
+| `web/gitea.go detailIssues` | `LinksByType` (gitea-repo) | /i/{path} issues card |
+| `web/gitea_writeback.go` | `GetByPath`, `LinksByType` | issue close/comment/create handlers |
+| `web/links.go` (add/remove/list) | `GetByPath`, `DatedLinks` | /i/{path} documents section |
+| `web/dashboard_pin.go` | `SetPinned` — WRITE, not slice B | pin toggle (slice C) |
+| `web/views.go handleViewRender` | `ListAll`, `AllTags`, `linkKindsByItem` | /views/{slug} render (5j) |
+| `web/system_views.go legacyRedirect` | `GetViewByID` — views CRUD (NOT in scope) | legacy 5i uuid → 5j slug redirect |
+| `internal/aggregate aggregator.go` | takes `LinkLister` interface (LinksByType + ItemsCreatedInRange) | shared fan-out across tasks/events/issues/docs |
+| `mcp/tools.go` (read tools) | `ListByFilters`, `LinksByRefType`, `GetByID`, `GetByPathOrSlug`, `LinksByType`, `ListAll`, `Search`, `RecentDocuments` (via dashboard fan-out reuse) | every read-side MCP tool |
+
+### §1.3 — Direct Pool access (out-of-scope for slice B, flagged for slice C)
+
+These bypass the store API and pull `*pgxpool.Pool` directly. Slice C (write-path) reworks them; flagging here so slice B's interface stays minimal:
+
+- `web/admin.go` — three count queries (`SELECT count(*) FROM projax.items WHERE …`) for the admin index. Either: (a) add `Counts(ctx) (AdminCounts, error)` to the adapter, (b) compute in-handler from `ListAll`. Adapter pick.
+- `web/bulk.go handleBulkApply` — multi-row UPDATE inside a tx. Pure write; slice C.
+- `web/links.go handleSetEventDate` — single UPDATE on `item_links.event_date`. Pure write; slice C.
+
+### §1.4 — `*Item` + `*ItemLink` shape contract (consumer side)
+
+Adapter MUST return these exact field sets in the result types. Nothing under `metadata.projax.*` in mBrian leaks to consumers; the adapter parses + materialises into the `Item` fields below.
+
+| field | semantics in slice B adapter |
+|---|---|
+| `Item.ID` | mBrian node uuid (post-migration); preserved old uuid OK per Q11 |
+| `Item.Kind` | `[]string{"project", ...}` — mBrian `node.type[]` 1:1 |
+| `Item.Title`, `Item.Slug`, `Item.ContentMD`, `Item.Aliases` | mBrian `node.title/slug/content_md/aliases` 1:1 |
+| `Item.Paths` | **derived** from `child_of` edge walk + the node's own slug. Adapter computes per-call (cached per-request) |
+| `Item.ParentIDs` | **derived** from outbound `child_of` edges |
+| `Item.Metadata` | `node.metadata` MINUS the `projax` sub-key (which gets unpacked into the struct fields below) |
+| `Item.Status` | `node.metadata.projax.status` (default "active") |
+| `Item.Pinned`, `Item.Archived` | `node.pinned`, `node.archived` 1:1 |
+| `Item.StartTime`, `Item.EndTime` | `node.metadata.projax.start_time` / `.end_time` (timestamptz strings) |
+| `Item.Tags`, `Item.Management`, `Item.TimelineExclude` | `node.metadata.projax.tags` / `.management` / `.timeline_exclude` |
+| `Item.Public`, `Item.PublicDescription`, `Item.PublicLiveURL`, `Item.PublicSourceURL`, `Item.PublicScreenshots` | `node.metadata.projax.public.{enabled, description, live_url, source_url, screenshots}` |
+| `Item.CreatedAt`, `Item.UpdatedAt` | `node.created_at`, `node.updated_at` 1:1 |
+| `Item.Source` | always `"projax"` (legacy field; new adapter sets this to maintain consumer assumption) |
+| `Item.SourceRefID` | mai.projects.id from `projax-mai-project` edge metadata when present |
+| `ItemLink.ID` | mBrian edge uuid |
+| `ItemLink.ItemID` | edge `source_id` (the projax-side end) |
+| `ItemLink.RefType` | strip `projax-` prefix from edge `rel` (`projax-caldav-list` → `caldav-list`) |
+| `ItemLink.RefID` | edge `metadata.ref_id` OR derived from rel-specific payload (caldav: `url`; gitea-repo: `owner/repo`; mai-project: `mai_project_id`) — see §3 gaps |
+| `ItemLink.Rel` | edge `note` (free-form annotation) OR a constant per rel-type (e.g. `'contains'`) |
+| `ItemLink.Metadata` | edge `metadata` MINUS the `ref_id` extraction |
+| `ItemLink.EventDate` | edge `metadata.event_date` (date string parsed) |
+| `ItemLink.CreatedAt` | edge `created_at` 1:1 |
+
+### §1.5 — Views (Phase 5j) — explicitly NOT in slice B
+
+Per m's Q5=(a), `projax.views` stays projax-resident. All view CRUD methods (`ListViews`, `GetView`, `GetViewByID`, `CreateView`, `UpdateView`, `DeleteView`, `TouchView`, `MostRecentView`, `ReorderViews`) stay on the existing `*Store` and are NOT part of the adapter interface. The `Server` struct uses the adapter for items+links and the existing `Store` for views.
+
+---
+
+## §2 — Adapter interface contract
+
+Defined in `store/adapter.go` (this branch). Pure projax-shaped structs in/out; zero mBrian type leakage. The existing `*store.Store` already satisfies this interface (it's just a subset of its public surface) — the compile-time assertion makes that explicit. Slice B impl ships a second satisfier (`*MBrianReader`) that wraps mBrian access.
+
+```go
+// ItemReader is the read-only contract every projax UI handler / aggregator /
+// MCP read tool depends on. Slice B implements a second satisfier on top of
+// mBrian's MCP/SQL surface.
+type ItemReader interface {
+    // Item lookups
+    ListAll(ctx context.Context) ([]*Item, error)
+    GetByID(ctx context.Context, id string) (*Item, error)
+    GetByPath(ctx context.Context, path string) (*Item, error)
+    GetByPathOrSlug(ctx context.Context, key string) (*Item, error)
+    Roots(ctx context.Context) ([]*Item, error)
+    MaiOrphans(ctx context.Context) ([]*Item, error)
+    ListByFilters(ctx context.Context, f SearchFilters) ([]*Item, error)
+    Search(ctx context.Context, q string, limit int) ([]*Item, error)
+    ItemsCreatedInRange(ctx context.Context, from, to time.Time) ([]*Item, error)
+    AllTags(ctx context.Context) ([]string, error)
+
+    // Link lookups
+    LinksByType(ctx context.Context, itemID, refType string) ([]*ItemLink, error)
+    LinksByRefType(ctx context.Context, refType string) ([]*ItemLink, error)
+    DatedLinks(ctx context.Context, itemID string) ([]*ItemLink, error)
+    DatedLinksRange(ctx context.Context, from, to time.Time) ([]*ItemLinkWithItem, error)
+    RecentDocuments(ctx context.Context, since time.Time, limit int) ([]*ItemLinkWithItem, error)
+}
+```
+
+### §2.1 — Methods needing edge-walk-derived data
+
+Slice B's mBrian impl must compute these from `child_of` edges + node fields. Cost is one outbound-edges fetch per node OR one bulk edges-by-rel query per request, depending on how the adapter caches.
+
+- `Item.Paths` — every method returning `*Item` or `[]*Item`.
+- `Item.ParentIDs` — same.
+- `GetByPath` — walks edges to resolve `dev.paliad` to a leaf node.
+- `Roots` — filter where no outbound `child_of` edge.
+- `MaiOrphans` — `Roots` ∩ `metadata.projax.management ⊇ {'mai'}`.
+
+### §2.2 — Methods needing metadata-unpack
+
+Adapter parses `metadata.projax.*` on read; writes (slice C) re-serialise. Affected fields: Status, Tags, Management, TimelineExclude, Public + 4 public_* fields, StartTime, EndTime.
+
+### §2.3 — Methods needing edge.metadata filters
+
+- `LinksByType(itemID, refType)`: WHERE source_id=$1 AND rel = 'projax-' || $2.
+- `LinksByRefType(refType)`: WHERE rel = 'projax-' || $1.
+- `DatedLinks(itemID)`: source_id=$1 AND metadata ? 'event_date'.
+- `DatedLinksRange(from, to)`: metadata->>'event_date' BETWEEN $1 AND $2.
+- `RecentDocuments(since, limit)`: dated links since $1 ORDER BY metadata->>'event_date' DESC LIMIT $2.
+
+mBrian's `idx_edges_metadata` GIN index already exists (mig 010); these queries are index-eligible.
+
+---
+
+## §3 — Gap flags
+
+Items the known mBrian schema needs to satisfy cleanly. The migration script handles most; flag here for the slice-B impl + the migration worker as cross-check items.
+
+| gap | shape | status |
+|---|---|---|
+| **`item_links.rel` (free-form annotation) preservation** | projax has both a typed `ref_type` AND a free-form `rel` text (`"contains"`, `"source"`, etc.) on item_links. mBrian's edge `rel` is the typed name; the free-form annotation maps to `edge.note`. Migration must NOT drop the projax `rel` value. | Add to m/mBrian#73 §1 edge mapping: source `rel` → mBrian `edges.note`. |
+| **`ItemLink.RefID` semantics per type** | projax `ref_id` is a typed external pointer (caldav url, gitea `owner/repo`, gitea-issue id, mai project uuid, bare url). mBrian edges carry the payload in metadata. Need a per-rel-type extraction rule. Suggested: `metadata.ref_id` for the canonical reference + leaves structured payload alongside (`url` for caldav, `owner`/`repo` for gitea). | Slice B impl reads back per-rel-type; document in m/mBrian#73 issue for the migration script to write consistently. |
+| **`paths text[]` recomputation cost** | Adapter computes paths from `child_of` edge walk per call. For `ListAll` over ~65 items, one bulk edges-by-rel query joined with the node id set is N rows where N = total `child_of` edges. Cheap at m's scale; add per-request memoisation. | Slice B impl. No mBrian-side action. |
+| **`AllTags` aggregation** | Union of `metadata.projax.tags[]` across all projax-managed nodes. No mBrian index on metadata-array-element. At m's scale (<200 nodes), full-scan is fine; if we grow, add a derived `[tag]` node graph (m's Q8 deferred to Phase 7). | Slice B impl, no mBrian-side action. |
+| **`Roots` / `MaiOrphans` predicate** | "No outbound `child_of` edge" requires a subquery / left-join-where-null pattern. Index-eligible via `idx_edges_source_rel` on `(source_id, rel)`. | Slice B impl. |
+| **`ItemsCreatedInRange`** | Direct over `nodes.created_at`; trivial. Scoped to `metadata.projax_origin IS NOT NULL` so non-projax mBrian nodes don't leak into projax surfaces. | Slice B impl + a `metadata GIN` query (already indexed). |
+| **`Item.Source` field expectation** | The legacy `Source` field on `Item` reads `"projax"` everywhere consumers check it (some MCP tools branch on it). Adapter sets a constant. | Slice B impl detail, no DB action. |
+| **`SourceRefID` for mai bridge** | When a node has a `projax-mai-project` edge, expose its `metadata.mai_project_id` as `Item.SourceRefID`. Slice D (mai bridge worker) writes these edges. | Slice B impl reads existing edges; slice D writes new ones. |
+| **`ItemLinkWithItem` join shape** | Used by `DatedLinksRange` and `RecentDocuments`. Adapter does two queries (edges-with-dates + node-by-id batch) + an in-memory join, OR one combined MCP call if mBrian exposes a bulk-edges-with-source-node helper. Both work; pick by perf. | Slice B impl, no mBrian-side change required. |
+| **Admin counts (web/admin.go direct Pool)** | Three count(*) queries (total items, total mai-managed, total public). Adapter gains `Counts(ctx) (AdminCounts, error)` — small extension. | Add to ItemReader interface in slice B (low-risk; constant-return until impl) OR keep as a separate `AdminReader` interface. Recommend adding to ItemReader for cohesion. |
+
+---
+
+## §4 — Skeleton (this branch)
+
+The Go file `store/adapter.go` ships in this branch with:
+
+1. `ItemReader` interface as in §2.
+2. `var _ ItemReader = (*Store)(nil)` compile-time assertion. (Drops in cleanly because `*Store` already exposes every method in the contract.)
+3. `MBrianReader` struct with stubbed method bodies that return `errNotImplementedSliceB`. Each stub carries a one-line comment naming the §3 gap it depends on (if any) so slice B's impl-fill knows what to look up.
+4. `var _ ItemReader = (*MBrianReader)(nil)` compile-time assertion so the stubs stay aligned with the interface.
+
+`go build ./...` is green with the skeleton in place. No tests, no behaviour, no mBrian client dependency.
+
+The actual mBrian client wiring (whether MCP-over-stdio, direct Postgres against `mbrian.*` schema, or the in-process submodule pattern flexsiebels uses) is the first decision slice-B-impl makes; it stays out of this prep step.
+
+---
+
+## §5 — Wiring shape after slice B impl
+
+For reference of the post-slice-B shape (no code in this slice):
+
+```go
+// Server struct keeps two readers: ItemReader (slice-B mBrian-backed) +
+// existing *Store (views CRUD only).
+type Server struct {
+    Items   ItemReader     // slice B: MBrianReader; today: *Store
+    Store   *store.Store   // views CRUD only after slice B
+    // ... rest unchanged
+}
+```
+
+Every handler that today reads `s.Store.ListAll(...)` becomes `s.Items.ListAll(...)`. Mechanical rename. Slice B impl ships both — adapter wiring + the rename across handlers — as one diff once the migration completes.
+
+---
+
+## §6 — What's NOT in this prep
+
+- mBrian-MCP client wiring.
+- Any test of mBrian-backed behaviour.
+- Write-path methods (slice C scope).
+- View CRUD migration (Q5=(a) stays projax-resident).
+- mai bridge worker (slice D).
+- Drop projax tables (slice E).
+
+---
+
+## §7 — References
+
+- `docs/plans/mbrian-backend-migration.md` (on `main`) — parent plan.
+- `cmd/projax-snapshot/` (slice 0, merged at `38182df`) — input for mBrian's migration.
+- m/mBrian#73 — mBrian-side schema convention node + migration script (in flight).
+- `store/store.go` — current `*Store` implementation; the interface `*Store` already satisfies.
+- `internal/aggregate/aggregator.go` — existing `LinkLister` interface precedent (a narrow projection of `*Store`).
--- a/docs/plans/slice-b-views-projectid-gap.md
+++ b/docs/plans/slice-b-views-projectid-gap.md
@@ -0,0 +1,103 @@
+# Slice B gap — `projax.views.filter_json.project_id` uuid-map
+
+**Status**: gap flagged + tooling shipped. Must run remap BEFORE Slice E
+drops `projax.items`.
+**Branch**: `mai/kahn/phase-6-sliceB`.
+**Author**: kahn (coder), 2026-05-31.
+**Parent**: `docs/plans/mbrian-backend-migration.md` (§7 lossy bits).
+
+## The gap
+
+Phase 5j saved views (`projax.views.filter_json`) carry `project_id` set
+to the **OLD** `projax.items.id` uuid. After Slice B's read path flips
+to mBrian, those uuids no longer resolve — a saved view scoped to
+`dev.paliad` references an id that mBrian's `paliad` node doesn't have
+(the migration script issued a fresh uuid + recorded the old id under
+`metadata.projax_origin` per m's Q11).
+
+Symptoms once `PROJAX_BACKEND=mbrian` rolls in production:
+- Opening `/views/{slug}` for any saved view that carries
+  `filter_json.project_id` returns the unfiltered set (the scope filter
+  silently no-ops).
+- The chip rendering still labels the view "scoped to dev.paliad"
+  because the cached `project_path` is unaffected — the discrepancy
+  surfaces as "label says scope, content doesn't filter."
+
+This must be fixed before Slice E drops `projax.items` or the old-id
+provenance disappears.
+
+## Two viable remediations
+
+### (a) One-shot uuid remap via the migration audit map
+
+mBrian dropped the `{old_projax_uuid → new_mbrian_uuid}` map at a shared
+mRiver path (head will relay). One-time SQL:
+
+```sql
+UPDATE projax.views
+SET filter_json = jsonb_set(
+    filter_json,
+    '{project_id}',
+    to_jsonb(<new>::text)
+)
+WHERE filter_json->>'project_id' = '<old>';
+```
+
+…iterated across the map. Pure data fix, no schema change. Idempotent
+(re-running against an already-remapped row is a no-op because the old
+uuid no longer matches).
+
+### (b) Resolve by slug instead of id
+
+Change the views resolver to look up `metadata.projax_origin = <id>`
+when the new uuid doesn't resolve, OR store the slug in
+`filter_json.project_path` (already present) as the canonical pointer.
+Heavier change in the views read path; ships in a future slice.
+
+## Recommended path
+
+**Hybrid**: (a) for the existing rows now (idempotent, surgical), and
+flag (b) as a follow-up in slice C/D for new view writes (use slug
+instead of uuid in the editor — the slug is the durable name post-
+migration anyway).
+
+## Tool shipped this slice
+
+`cmd/projax-remap-views/main.go` (this commit). Usage:
+
+```
+projax-remap-views --map /path/to/uuid-map.json
+```
+
+Map shape:
+```json
+{
+  "old-projax-uuid-1": "new-mbrian-uuid-1",
+  "old-projax-uuid-2": "new-mbrian-uuid-2",
+  ...
+}
+```
+
+The tool:
+- Loads the uuid map from JSON.
+- Walks every live `projax.views` row.
+- For each row with `filter_json.project_id` matching a map key,
+  rewrites it to the new uuid.
+- Prints a per-row before/after summary; commit only happens on
+  `--apply`. Default is dry-run.
+
+Idempotent on re-run: rows already pointing at a new uuid don't match
+any map key on the second pass, so they pass through untouched.
+
+## When to run
+
+After the mBrian migration completes (already shipped) + after head
+relays the uuid-map path. Before any user is on
+`PROJAX_BACKEND=mbrian`. The tool runs against msupabase using the same
+`SUPABASE_DATABASE_URL` the projax binary uses.
+
+## Validation
+
+Post-remap: pick the spot-check saved views (if any exist), open them
+through both `PROJAX_BACKEND=store` and `PROJAX_BACKEND=mbrian`, confirm
+the filtered set matches.
--- a/store/adapter.go
+++ b/store/adapter.go
@@ -0,0 +1,52 @@
+package store
+
+import (
+	"context"
+	"time"
+)
+
+// ItemReader is the read-path contract every projax UI handler, the
+// internal/aggregate fan-out engine, and the MCP read tools depend on.
+// Pure projax-shaped structs in/out; the slice-B mBrian-backed
+// implementation translates mBrian nodes/edges into the same shape
+// without leaking mBrian types to consumers.
+//
+// Phase 6 Slice B (live impl) — see store/mbrian.go for the MBrianReader
+// implementation against the migrated mbrian.* schema, and
+// docs/plans/slice-b-adapter-contract.md for the consumer inventory +
+// per-method semantics.
+//
+// Two satisfiers ship:
+//   *Store        — pgx-backed against projax.items (today; legacy).
+//   *MBrianReader — pgx-backed against mbrian.{nodes,edges} (slice B).
+//
+// Selection between them is wired at Server-construction time via
+// PROJAX_BACKEND=store|mbrian (defaults to "store" until slice B is
+// rolled to production).
+type ItemReader interface {
+	// --- item lookups ---
+	ListAll(ctx context.Context) ([]*Item, error)
+	GetByID(ctx context.Context, id string) (*Item, error)
+	GetByPath(ctx context.Context, path string) (*Item, error)
+	GetByPathOrSlug(ctx context.Context, key string) (*Item, error)
+	Roots(ctx context.Context) ([]*Item, error)
+	MaiOrphans(ctx context.Context) ([]*Item, error)
+	ListByFilters(ctx context.Context, f SearchFilters) ([]*Item, error)
+	Search(ctx context.Context, q string, limit int) ([]*Item, error)
+	ItemsCreatedInRange(ctx context.Context, from, to time.Time) ([]*Item, error)
+	AllTags(ctx context.Context) ([]string, error)
+
+	// --- link lookups ---
+	LinksByType(ctx context.Context, itemID, refType string) ([]*ItemLink, error)
+	LinksByRefType(ctx context.Context, refType string) ([]*ItemLink, error)
+	DatedLinks(ctx context.Context, itemID string) ([]*ItemLink, error)
+	DatedLinksRange(ctx context.Context, from, to time.Time) ([]*ItemLinkWithItem, error)
+	RecentDocuments(ctx context.Context, since time.Time, limit int) ([]*ItemLinkWithItem, error)
+}
+
+// Compile-time assertion that the existing pgx-backed *Store satisfies
+// ItemReader. Drops in cleanly because every method in the interface is
+// already part of *Store's public surface. If a future refactor removes
+// or reshapes one of these methods on *Store, the compiler points at
+// this line first.
+var _ ItemReader = (*Store)(nil)
--- a/store/mbrian.go
+++ b/store/mbrian.go
@@ -0,0 +1,990 @@
+package store
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"sort"
+	"strings"
+	"time"
+
+	"github.com/jackc/pgx/v5"
+	"github.com/jackc/pgx/v5/pgxpool"
+)
+
+// Phase 6 Slice B — MBrianReader is the live read-path adapter against
+// the migrated mBrian graph (msupabase, schema `mbrian`). Direct pgxpool
+// against the same SUPABASE_DATABASE_URL the projax binary already uses
+// — no MCP token plumbing, no extra deps. Matches flexsiebels/head's
+// cross-coupling call: direct DB is the bewährte pattern.
+//
+// Mapping contract (see docs/plans/slice-b-adapter-contract.md):
+//   * projax-managed nodes are mbrian.nodes where metadata ? 'projax_origin'.
+//   * Item.Paths + Item.ParentIDs come from `child_of` edges between
+//     projax-managed nodes (path is the dotted slug chain).
+//   * Item.Status / .Tags / .Management / .Public* / .StartTime /
+//     .EndTime / .TimelineExclude unpack from metadata.projax.*.
+//   * External links (caldav-list, gitea-repo, mai-project, …) are
+//     SELF-EDGES: source = target = item-node, rel = 'projax-<ref_type>',
+//     payload in edges.metadata.
+//   * Item.CreatedAt / .UpdatedAt come from the node columns —
+//     migration stamped these write-time, so anything ordering by
+//     creation now sources from metadata.projax.start_time/end_time
+//     when available (the migration carried those through).
+
+// MBrianReader is the slice-B read-path adapter. Wraps a pgxpool that
+// reaches the mbrian.* schema (same SUPABASE_DATABASE_URL the projax
+// binary uses for projax.*).
+type MBrianReader struct {
+	pool *pgxpool.Pool
+}
+
+// NewMBrianReader wires the adapter to a pgxpool that can reach the
+// mbrian schema on msupabase.
+func NewMBrianReader(pool *pgxpool.Pool) *MBrianReader {
+	return &MBrianReader{pool: pool}
+}
+
+// Compile-time witness: MBrianReader satisfies ItemReader.
+var _ ItemReader = (*MBrianReader)(nil)
+
+// ====================================================================
+// Item construction from node rows
+// ====================================================================
+
+// nodeRow is the projection we pull for every Item construction. Matches
+// the SELECT in itemQuery below.
+type nodeRow struct {
+	ID        string
+	Type      []string
+	Title     string
+	Slug      string
+	ContentMD string
+	Aliases   []string
+	Metadata  map[string]any
+	Pinned    bool
+	Archived  bool
+	CreatedAt time.Time
+	UpdatedAt time.Time
+}
+
+const projaxNodeColumns = `n.id::text, n.type, n.title, n.slug, n.content_md,
+       n.aliases, n.metadata, n.pinned, n.archived,
+       n.created_at, n.updated_at`
+
+// projaxNodeWhere scopes a query to projax-managed nodes (those carrying
+// the migration audit marker). Live + non-deleted only.
+const projaxNodeWhere = `n.deleted_at IS NULL AND n.metadata ? 'projax_origin'`
+
+// scanNodeRow consumes the projection above.
+func scanNodeRow(s interface {
+	Scan(dest ...any) error
+}) (*nodeRow, error) {
+	r := &nodeRow{}
+	if err := s.Scan(&r.ID, &r.Type, &r.Title, &r.Slug, &r.ContentMD,
+		&r.Aliases, &r.Metadata, &r.Pinned, &r.Archived,
+		&r.CreatedAt, &r.UpdatedAt); err != nil {
+		return nil, err
+	}
+	if r.Type == nil {
+		r.Type = []string{}
+	}
+	if r.Aliases == nil {
+		r.Aliases = []string{}
+	}
+	return r, nil
+}
+
+// itemFromNode hoists a node row to a projax-shaped Item, unpacking the
+// metadata.projax.* fields. Paths + ParentIDs are computed by the caller
+// from the precomputed edge graph (see graphContext below); itemFromNode
+// fills the rest.
+func itemFromNode(r *nodeRow) *Item {
+	it := &Item{
+		ID:        r.ID,
+		Kind:      r.Type,
+		Title:     r.Title,
+		Slug:      r.Slug,
+		ContentMD: r.ContentMD,
+		Aliases:   r.Aliases,
+		Pinned:    r.Pinned,
+		Archived:  r.Archived,
+		CreatedAt: r.CreatedAt,
+		UpdatedAt: r.UpdatedAt,
+		Source:    "projax",
+		Status:    "active", // default if not in metadata.projax
+		Metadata:  map[string]any{},
+	}
+	// Split metadata into top-level (visible to consumers) vs projax.* (unpacked).
+	projaxMeta := map[string]any{}
+	for k, v := range r.Metadata {
+		switch k {
+		case "projax":
+			if m, ok := v.(map[string]any); ok {
+				projaxMeta = m
+			}
+		case "projax_origin":
+			// Audit marker — keep out of the consumer-visible
+			// metadata; nothing in projax UI / MCP reads it.
+		default:
+			it.Metadata[k] = v
+		}
+	}
+	if v, ok := projaxMeta["status"].(string); ok && v != "" {
+		it.Status = v
+	}
+	if v, ok := projaxMeta["tags"]; ok {
+		it.Tags = anyToStringSlice(v)
+	}
+	if v, ok := projaxMeta["management"]; ok {
+		it.Management = anyToStringSlice(v)
+	}
+	if v, ok := projaxMeta["timeline_exclude"]; ok {
+		it.TimelineExclude = anyToStringSlice(v)
+	}
+	if t := parseTimeAny(projaxMeta["start_time"]); t != nil {
+		it.StartTime = t
+	}
+	if t := parseTimeAny(projaxMeta["end_time"]); t != nil {
+		it.EndTime = t
+	}
+	if pub, ok := projaxMeta["public"].(map[string]any); ok {
+		if v, ok := pub["enabled"].(bool); ok {
+			it.Public = v
+		}
+		it.PublicDescription, _ = pub["description"].(string)
+		it.PublicLiveURL, _ = pub["live_url"].(string)
+		it.PublicSourceURL, _ = pub["source_url"].(string)
+		it.PublicScreenshots = anyToStringSlice(pub["screenshots"])
+	}
+	if it.Tags == nil {
+		it.Tags = []string{}
+	}
+	if it.Management == nil {
+		it.Management = []string{}
+	}
+	if it.TimelineExclude == nil {
+		it.TimelineExclude = []string{}
+	}
+	if it.PublicScreenshots == nil {
+		it.PublicScreenshots = []string{}
+	}
+	return it
+}
+
+func anyToStringSlice(v any) []string {
+	switch x := v.(type) {
+	case []string:
+		return x
+	case []any:
+		out := make([]string, 0, len(x))
+		for _, e := range x {
+			if s, ok := e.(string); ok {
+				out = append(out, s)
+			}
+		}
+		return out
+	}
+	return nil
+}
+
+func parseTimeAny(v any) *time.Time {
+	s, ok := v.(string)
+	if !ok || s == "" {
+		return nil
+	}
+	for _, layout := range []string{time.RFC3339, time.RFC3339Nano, "2006-01-02T15:04:05Z", "2006-01-02"} {
+		if t, err := time.Parse(layout, s); err == nil {
+			return &t
+		}
+	}
+	return nil
+}
+
+// ====================================================================
+// graphContext — bulk edge fetch reused across read methods
+// ====================================================================
+
+// graphContext caches the projax-edge graph + node-id lookups for one
+// adapter call. ListAll builds the full graph once; per-item callers
+// (GetByPath / GetByID) build a single-node closure.
+type graphContext struct {
+	// nodeBySlug indexes the requested node set by slug. Used by path
+	// resolution.
+	nodeBySlug map[string]*nodeRow
+	// nodeByID indexes by uuid. Used by parent / outbound traversal.
+	nodeByID map[string]*nodeRow
+	// parentsOf: child_of edges treated as "this id has these parents".
+	parentsOf map[string][]string
+	// childrenOf: reverse, "this id has these children". Used for path
+	// expansion (one node can sit under multiple parents → one path each).
+	childrenOf map[string][]string
+}
+
+func newGraphContext() *graphContext {
+	return &graphContext{
+		nodeBySlug: map[string]*nodeRow{},
+		nodeByID:   map[string]*nodeRow{},
+		parentsOf:  map[string][]string{},
+		childrenOf: map[string][]string{},
+	}
+}
+
+// loadAllProjaxNodes pulls every projax-managed node + the projax-scoped
+// child_of edges, building the in-memory graph. Two queries — cheap at
+// m's scale (~65 nodes, ~78 edges).
+func loadAllProjaxNodes(ctx context.Context, pool *pgxpool.Pool) (*graphContext, error) {
+	gc := newGraphContext()
+	nrows, err := pool.Query(ctx,
+		`SELECT `+projaxNodeColumns+`
+FROM mbrian.nodes n
+WHERE `+projaxNodeWhere+`
+ORDER BY n.slug`)
+	if err != nil {
+		return nil, fmt.Errorf("query nodes: %w", err)
+	}
+	defer nrows.Close()
+	for nrows.Next() {
+		r, err := scanNodeRow(nrows)
+		if err != nil {
+			return nil, err
+		}
+		gc.nodeByID[r.ID] = r
+		gc.nodeBySlug[r.Slug] = r
+	}
+	if err := nrows.Err(); err != nil {
+		return nil, err
+	}
+	erows, err := pool.Query(ctx,
+		`SELECT e.source_id::text, e.target_id::text
+FROM mbrian.edges e
+WHERE e.rel = 'child_of'
+  AND e.source_id IN (SELECT id FROM mbrian.nodes WHERE metadata ? 'projax_origin' AND deleted_at IS NULL)
+  AND e.target_id IN (SELECT id FROM mbrian.nodes WHERE metadata ? 'projax_origin' AND deleted_at IS NULL)`)
+	if err != nil {
+		return nil, fmt.Errorf("query edges: %w", err)
+	}
+	defer erows.Close()
+	for erows.Next() {
+		var src, tgt string
+		if err := erows.Scan(&src, &tgt); err != nil {
+			return nil, err
+		}
+		gc.parentsOf[src] = append(gc.parentsOf[src], tgt)
+		gc.childrenOf[tgt] = append(gc.childrenOf[tgt], src)
+	}
+	return gc, erows.Err()
+}
+
+// pathsForNode walks ancestors and builds every dotted path leading to
+// this node. Multi-parent → multiple paths; mirrors the projax.items
+// `paths text[]` shape.
+//
+// Sorted + deduped output. Recursion depth-capped at 64 hops to match
+// projax's path trigger.
+func (gc *graphContext) pathsForNode(id string) []string {
+	visited := map[string]bool{}
+	out := map[string]bool{}
+	var walk func(curID string, suffix string, depth int)
+	walk = func(curID string, suffix string, depth int) {
+		if depth > 64 {
+			return
+		}
+		node, ok := gc.nodeByID[curID]
+		if !ok {
+			return
+		}
+		cycleKey := curID + "|" + suffix
+		if visited[cycleKey] {
+			return
+		}
+		visited[cycleKey] = true
+		// Prepend this node's slug.
+		var here string
+		if suffix == "" {
+			here = node.Slug
+		} else {
+			here = node.Slug + "." + suffix
+		}
+		parents := gc.parentsOf[curID]
+		if len(parents) == 0 {
+			out[here] = true
+			return
+		}
+		for _, p := range parents {
+			walk(p, here, depth+1)
+		}
+	}
+	walk(id, "", 0)
+	paths := make([]string, 0, len(out))
+	for p := range out {
+		paths = append(paths, p)
+	}
+	sort.Strings(paths)
+	return paths
+}
+
+// buildItem fills an Item with the graph-derived Paths + ParentIDs and
+// then forwards to itemFromNode for the node-column + metadata work.
+func (gc *graphContext) buildItem(id string) *Item {
+	r, ok := gc.nodeByID[id]
+	if !ok {
+		return nil
+	}
+	it := itemFromNode(r)
+	it.Paths = gc.pathsForNode(id)
+	parents := gc.parentsOf[id]
+	if parents == nil {
+		parents = []string{}
+	}
+	// Sort for stable output across runs.
+	sort.Strings(parents)
+	it.ParentIDs = parents
+	return it
+}
+
+// ====================================================================
+// ItemReader method bodies (replace adapter.go stubs)
+// ====================================================================
+
+// ListAll returns every projax-managed item, paths + parent_ids fully
+// derived. One graph build per call; cheap at m's scale.
+func (r *MBrianReader) ListAll(ctx context.Context) ([]*Item, error) {
+	gc, err := loadAllProjaxNodes(ctx, r.pool)
+	if err != nil {
+		return nil, err
+	}
+	out := make([]*Item, 0, len(gc.nodeByID))
+	for id := range gc.nodeByID {
+		out = append(out, gc.buildItem(id))
+	}
+	sort.Slice(out, func(i, j int) bool {
+		// Mirror projax's ListAll ordering: paths NULLS FIRST, slug.
+		ip, jp := out[i].PrimaryPath(), out[j].PrimaryPath()
+		if ip == jp {
+			return out[i].Slug < out[j].Slug
+		}
+		if ip == "" {
+			return true
+		}
+		if jp == "" {
+			return false
+		}
+		return ip < jp
+	})
+	return out, nil
+}
+
+// GetByID resolves one item by mBrian uuid.
+func (r *MBrianReader) GetByID(ctx context.Context, id string) (*Item, error) {
+	gc, err := loadAllProjaxNodes(ctx, r.pool)
+	if err != nil {
+		return nil, err
+	}
+	if _, ok := gc.nodeByID[id]; !ok {
+		return nil, ErrNotFound
+	}
+	return gc.buildItem(id), nil
+}
+
+// GetByPath resolves a dotted path (`dev.paliad`, `work.upc.deadlines`)
+// to its leaf node, then materialises via the graph context.
+func (r *MBrianReader) GetByPath(ctx context.Context, path string) (*Item, error) {
+	if path == "" {
+		return nil, ErrNotFound
+	}
+	gc, err := loadAllProjaxNodes(ctx, r.pool)
+	if err != nil {
+		return nil, err
+	}
+	parts := strings.Split(path, ".")
+	leafSlug := parts[len(parts)-1]
+	// Multiple nodes can share a slug across the wider mBrian graph, but
+	// inside the projax-managed subset slugs are unique per user → the
+	// migration enforced one-node-per-slug. Pick that node.
+	node, ok := gc.nodeBySlug[leafSlug]
+	if !ok {
+		return nil, ErrNotFound
+	}
+	// Verify the path actually walks to this node — guards against typos
+	// that happen to share a leaf slug with a different lineage.
+	paths := gc.pathsForNode(node.ID)
+	for _, p := range paths {
+		if p == path {
+			return gc.buildItem(node.ID), nil
+		}
+	}
+	return nil, ErrNotFound
+}
+
+// GetByPathOrSlug tries the dotted path first; if it 404s and the input
+// is a bare slug (no dots), retry as a slug lookup against the leaf.
+func (r *MBrianReader) GetByPathOrSlug(ctx context.Context, key string) (*Item, error) {
+	if it, err := r.GetByPath(ctx, key); err == nil {
+		return it, nil
+	} else if !errors.Is(err, ErrNotFound) {
+		return nil, err
+	}
+	// Bare slug fallback.
+	if strings.Contains(key, ".") {
+		return nil, ErrNotFound
+	}
+	gc, err := loadAllProjaxNodes(ctx, r.pool)
+	if err != nil {
+		return nil, err
+	}
+	if node, ok := gc.nodeBySlug[key]; ok {
+		return gc.buildItem(node.ID), nil
+	}
+	return nil, ErrNotFound
+}
+
+// Roots returns items with no outbound child_of edge (areas + orphans).
+func (r *MBrianReader) Roots(ctx context.Context) ([]*Item, error) {
+	gc, err := loadAllProjaxNodes(ctx, r.pool)
+	if err != nil {
+		return nil, err
+	}
+	out := []*Item{}
+	for id := range gc.nodeByID {
+		if len(gc.parentsOf[id]) == 0 {
+			out = append(out, gc.buildItem(id))
+		}
+	}
+	sort.Slice(out, func(i, j int) bool { return out[i].Slug < out[j].Slug })
+	return out, nil
+}
+
+// MaiOrphans returns root mai-managed items needing classification —
+// projax's /admin/classify surface.
+func (r *MBrianReader) MaiOrphans(ctx context.Context) ([]*Item, error) {
+	gc, err := loadAllProjaxNodes(ctx, r.pool)
+	if err != nil {
+		return nil, err
+	}
+	out := []*Item{}
+	for id, n := range gc.nodeByID {
+		if len(gc.parentsOf[id]) > 0 {
+			continue
+		}
+		it := gc.buildItem(id)
+		if it == nil {
+			continue
+		}
+		if !it.HasManagement("mai") {
+			continue
+		}
+		_ = n
+		out = append(out, it)
+	}
+	sort.Slice(out, func(i, j int) bool { return out[i].Slug < out[j].Slug })
+	return out, nil
+}
+
+// ListByFilters filters in-memory after a full graph load. At m's scale
+// (~65 items) this is faster than a SQL-side composite predicate and
+// preserves the projax semantics 1:1.
+func (r *MBrianReader) ListByFilters(ctx context.Context, f SearchFilters) ([]*Item, error) {
+	items, err := r.ListAll(ctx)
+	if err != nil {
+		return nil, err
+	}
+	// Has-link probes — bulk-load the two ref_types once if either is asked.
+	var hasRepo, hasCal map[string]bool
+	if f.HasRepo != nil {
+		hasRepo = map[string]bool{}
+		links, err := r.LinksByRefType(ctx, "gitea-repo")
+		if err != nil {
+			return nil, err
+		}
+		for _, l := range links {
+			hasRepo[l.ItemID] = true
+		}
+	}
+	if f.HasCalDAV != nil {
+		hasCal = map[string]bool{}
+		links, err := r.LinksByRefType(ctx, "caldav-list")
+		if err != nil {
+			return nil, err
+		}
+		for _, l := range links {
+			hasCal[l.ItemID] = true
+		}
+	}
+	out := []*Item{}
+	for _, it := range items {
+		if f.ParentPath != "" {
+			scoped := false
+			pfx := f.ParentPath + "."
+			for _, p := range it.Paths {
+				if p == f.ParentPath || strings.HasPrefix(p, pfx) {
+					scoped = true
+					break
+				}
+			}
+			if !scoped {
+				continue
+			}
+		}
+		if len(f.Tags) > 0 {
+			ok := true
+			for _, t := range f.Tags {
+				if !it.HasTag(t) {
+					ok = false
+					break
+				}
+			}
+			if !ok {
+				continue
+			}
+		}
+		if len(f.Management) > 0 {
+			ok := true
+			for _, m := range f.Management {
+				if !it.HasManagement(m) {
+					ok = false
+					break
+				}
+			}
+			if !ok {
+				continue
+			}
+		}
+		if len(f.Kind) > 0 {
+			ok := false
+			for _, k := range f.Kind {
+				if containsString(it.Kind, k) {
+					ok = true
+					break
+				}
+			}
+			if !ok {
+				continue
+			}
+		}
+		if f.Status != "" && it.Status != f.Status {
+			continue
+		}
+		if f.Q != "" && !itemMatchesSubstring(it, strings.ToLower(f.Q)) {
+			continue
+		}
+		if f.HasRepo != nil && hasRepo[it.ID] != *f.HasRepo {
+			continue
+		}
+		if f.HasCalDAV != nil && hasCal[it.ID] != *f.HasCalDAV {
+			continue
+		}
+		if f.Public != nil && it.Public != *f.Public {
+			continue
+		}
+		out = append(out, it)
+	}
+	if f.Limit > 0 && len(out) > f.Limit {
+		out = out[:f.Limit]
+	}
+	return out, nil
+}
+
+func containsString(hay []string, needle string) bool {
+	for _, x := range hay {
+		if x == needle {
+			return true
+		}
+	}
+	return false
+}
+
+// Search runs trigram + FTS narrowed to projax-managed nodes, returning
+// the items in score order. Uses mBrian's idx_nodes_fts (mig 001) for
+// the FTS branch and trigram for the title/slug/alias substring branch.
+func (r *MBrianReader) Search(ctx context.Context, q string, limit int) ([]*Item, error) {
+	q = strings.TrimSpace(q)
+	if q == "" {
+		return nil, nil
+	}
+	if limit <= 0 || limit > 200 {
+		limit = 50
+	}
+	gc, err := loadAllProjaxNodes(ctx, r.pool)
+	if err != nil {
+		return nil, err
+	}
+	// In-memory filter — m's scale doesn't justify a custom SQL ranking.
+	// Mirror Search behaviour from store.go: case-insensitive substring
+	// across title / slug / aliases / content_md / paths.
+	ql := strings.ToLower(q)
+	out := []*Item{}
+	for id := range gc.nodeByID {
+		it := gc.buildItem(id)
+		if it == nil {
+			continue
+		}
+		if itemMatchesSubstring(it, ql) {
+			out = append(out, it)
+		}
+		if len(out) >= limit {
+			break
+		}
+	}
+	sort.Slice(out, func(i, j int) bool { return out[i].Slug < out[j].Slug })
+	return out, nil
+}
+
+func itemMatchesSubstring(it *Item, q string) bool {
+	if strings.Contains(strings.ToLower(it.Title), q) {
+		return true
+	}
+	if strings.Contains(strings.ToLower(it.Slug), q) {
+		return true
+	}
+	if strings.Contains(strings.ToLower(it.ContentMD), q) {
+		return true
+	}
+	for _, a := range it.Aliases {
+		if strings.Contains(strings.ToLower(a), q) {
+			return true
+		}
+	}
+	for _, p := range it.Paths {
+		if strings.Contains(strings.ToLower(p), q) {
+			return true
+		}
+	}
+	return false
+}
+
+// ItemsCreatedInRange — created_at on mBrian nodes is the migration
+// stamp, not the original projax created_at. Order off
+// metadata.projax.start_time when present, fall back to created_at.
+func (r *MBrianReader) ItemsCreatedInRange(ctx context.Context, from, to time.Time) ([]*Item, error) {
+	items, err := r.ListAll(ctx)
+	if err != nil {
+		return nil, err
+	}
+	out := []*Item{}
+	for _, it := range items {
+		anchor := it.CreatedAt
+		if it.StartTime != nil {
+			anchor = *it.StartTime
+		}
+		if !anchor.Before(from) && anchor.Before(to) {
+			out = append(out, it)
+		}
+	}
+	sort.Slice(out, func(i, j int) bool {
+		a, b := out[i].CreatedAt, out[j].CreatedAt
+		if out[i].StartTime != nil {
+			a = *out[i].StartTime
+		}
+		if out[j].StartTime != nil {
+			b = *out[j].StartTime
+		}
+		return a.Before(b)
+	})
+	return out, nil
+}
+
+// AllTags unions metadata.projax.tags across every projax-managed node.
+// Full-scan; cheap at m's scale per §3 gap notes.
+func (r *MBrianReader) AllTags(ctx context.Context) ([]string, error) {
+	gc, err := loadAllProjaxNodes(ctx, r.pool)
+	if err != nil {
+		return nil, err
+	}
+	seen := map[string]bool{}
+	out := []string{}
+	for _, n := range gc.nodeByID {
+		if pm, ok := n.Metadata["projax"].(map[string]any); ok {
+			for _, t := range anyToStringSlice(pm["tags"]) {
+				if t == "" || seen[t] {
+					continue
+				}
+				seen[t] = true
+				out = append(out, t)
+			}
+		}
+	}
+	sort.Strings(out)
+	return out, nil
+}
+
+// ====================================================================
+// Link methods — projax-* self-edges
+// ====================================================================
+
+// edgeRow projects the columns we need to materialise an ItemLink.
+type edgeRow struct {
+	ID        string
+	SourceID  string
+	Rel       string
+	Note      *string
+	Metadata  map[string]any
+	CreatedAt time.Time
+}
+
+func scanEdgeRow(s interface {
+	Scan(dest ...any) error
+}) (*edgeRow, error) {
+	r := &edgeRow{}
+	if err := s.Scan(&r.ID, &r.SourceID, &r.Rel, &r.Note, &r.Metadata, &r.CreatedAt); err != nil {
+		return nil, err
+	}
+	if r.Metadata == nil {
+		r.Metadata = map[string]any{}
+	}
+	return r, nil
+}
+
+const edgeColumns = `e.id::text, e.source_id::text, e.rel, e.note, e.metadata, e.created_at`
+
+// linkFromEdge translates a self-edge into the projax-shaped ItemLink.
+// Per contract: ref_type = strip "projax-" prefix; ref_id derived from
+// edge.metadata per ref_type per the m/mBrian#73 contract.
+func linkFromEdge(r *edgeRow) *ItemLink {
+	refType := strings.TrimPrefix(r.Rel, "projax-")
+	l := &ItemLink{
+		ID:        r.ID,
+		ItemID:    r.SourceID,
+		RefType:   refType,
+		Rel:       "",
+		Metadata:  map[string]any{},
+		CreatedAt: r.CreatedAt,
+	}
+	// The original projax.item_links.rel (free-form annotation) lives at
+	// metadata.projax_rel — set it back on Rel.
+	if v, ok := r.Metadata["projax_rel"].(string); ok {
+		l.Rel = v
+	}
+	// Per-ref_type RefID extraction.
+	switch refType {
+	case "caldav-list":
+		if v, ok := r.Metadata["url"].(string); ok {
+			l.RefID = v
+		}
+	case "gitea-repo":
+		owner, _ := r.Metadata["owner"].(string)
+		repo, _ := r.Metadata["repo"].(string)
+		if owner != "" && repo != "" {
+			l.RefID = owner + "/" + repo
+		}
+	case "gitea-issue":
+		owner, _ := r.Metadata["owner"].(string)
+		repo, _ := r.Metadata["repo"].(string)
+		num, _ := r.Metadata["number"].(float64)
+		if owner != "" && repo != "" && num > 0 {
+			l.RefID = fmt.Sprintf("%s/%s#%d", owner, repo, int(num))
+		}
+	case "mai-project":
+		if v, ok := r.Metadata["mai_project_id"].(string); ok {
+			l.RefID = v
+		}
+	case "url", "doc", "document", "note":
+		if v, ok := r.Metadata["url"].(string); ok {
+			l.RefID = v
+		} else if v, ok := r.Metadata["path"].(string); ok {
+			l.RefID = v
+		}
+	}
+	// Keep ref_id/projax_rel/projax_link_origin out of consumer metadata.
+	for k, v := range r.Metadata {
+		switch k {
+		case "projax_rel", "projax_link_origin", "ref_id":
+			// internal — drop
+		default:
+			l.Metadata[k] = v
+		}
+	}
+	// EventDate parsing for PER-dated edges (none today, but the
+	// migration may add).
+	l.EventDate = parseTimeAny(r.Metadata["event_date"])
+	if r.Note != nil && *r.Note != "" {
+		l.Note = r.Note
+	}
+	return l
+}
+
+// LinksByType — one item's projax-* edges of a given ref_type. Empty
+// refType returns every projax-* edge for the item (matches store.go).
+func (r *MBrianReader) LinksByType(ctx context.Context, itemID, refType string) ([]*ItemLink, error) {
+	var rows pgx.Rows
+	var err error
+	if refType == "" {
+		rows, err = r.pool.Query(ctx,
+			`SELECT `+edgeColumns+`
+FROM mbrian.edges e
+WHERE e.source_id = $1 AND e.rel LIKE 'projax-%'
+ORDER BY e.created_at`, itemID)
+	} else {
+		rows, err = r.pool.Query(ctx,
+			`SELECT `+edgeColumns+`
+FROM mbrian.edges e
+WHERE e.source_id = $1 AND e.rel = $2
+ORDER BY e.created_at`, itemID, "projax-"+refType)
+	}
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	out := []*ItemLink{}
+	for rows.Next() {
+		er, err := scanEdgeRow(rows)
+		if err != nil {
+			return nil, err
+		}
+		out = append(out, linkFromEdge(er))
+	}
+	return out, rows.Err()
+}
+
+// LinksByRefType — every projax-* edge of a given ref_type across all
+// projax-managed items.
+func (r *MBrianReader) LinksByRefType(ctx context.Context, refType string) ([]*ItemLink, error) {
+	rows, err := r.pool.Query(ctx,
+		`SELECT `+edgeColumns+`
+FROM mbrian.edges e
+WHERE e.rel = $1
+ORDER BY e.created_at`, "projax-"+refType)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	out := []*ItemLink{}
+	for rows.Next() {
+		er, err := scanEdgeRow(rows)
+		if err != nil {
+			return nil, err
+		}
+		out = append(out, linkFromEdge(er))
+	}
+	return out, rows.Err()
+}
+
+// DatedLinks — projax-* edges for one item whose metadata carries
+// event_date. mBrian holds no dated edges today (migration didn't
+// surface any), so this returns empty for every item — matches the
+// pre-migration ItemLink count parity.
+func (r *MBrianReader) DatedLinks(ctx context.Context, itemID string) ([]*ItemLink, error) {
+	rows, err := r.pool.Query(ctx,
+		`SELECT `+edgeColumns+`
+FROM mbrian.edges e
+WHERE e.source_id = $1 AND e.rel LIKE 'projax-%' AND e.metadata ? 'event_date'
+ORDER BY e.metadata->>'event_date'`, itemID)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	out := []*ItemLink{}
+	for rows.Next() {
+		er, err := scanEdgeRow(rows)
+		if err != nil {
+			return nil, err
+		}
+		out = append(out, linkFromEdge(er))
+	}
+	return out, rows.Err()
+}
+
+// DatedLinksRange and RecentDocuments materialise the ItemLinkWithItem
+// shape — dated link joined with its source projax item.
+func (r *MBrianReader) DatedLinksRange(ctx context.Context, from, to time.Time) ([]*ItemLinkWithItem, error) {
+	return r.datedJoin(ctx,
+		`SELECT `+edgeColumns+`
+FROM mbrian.edges e
+WHERE e.rel LIKE 'projax-%' AND e.metadata ? 'event_date'
+  AND (e.metadata->>'event_date')::date BETWEEN $1 AND $2
+ORDER BY e.metadata->>'event_date' DESC`, from, to)
+}
+
+func (r *MBrianReader) RecentDocuments(ctx context.Context, since time.Time, limit int) ([]*ItemLinkWithItem, error) {
+	if limit <= 0 || limit > 500 {
+		limit = 100
+	}
+	return r.datedJoinLimit(ctx,
+		`SELECT `+edgeColumns+`
+FROM mbrian.edges e
+WHERE e.rel LIKE 'projax-%' AND e.metadata ? 'event_date'
+  AND (e.metadata->>'event_date')::date >= $1
+ORDER BY e.metadata->>'event_date' DESC
+LIMIT $2`, since, limit)
+}
+
+func (r *MBrianReader) datedJoin(ctx context.Context, sql string, from, to time.Time) ([]*ItemLinkWithItem, error) {
+	rows, err := r.pool.Query(ctx, sql, from.Format("2006-01-02"), to.Format("2006-01-02"))
+	if err != nil {
+		return nil, err
+	}
+	return r.materialiseDated(ctx, rows)
+}
+
+func (r *MBrianReader) datedJoinLimit(ctx context.Context, sql string, since time.Time, limit int) ([]*ItemLinkWithItem, error) {
+	rows, err := r.pool.Query(ctx, sql, since.Format("2006-01-02"), limit)
+	if err != nil {
+		return nil, err
+	}
+	return r.materialiseDated(ctx, rows)
+}
+
+func (r *MBrianReader) materialiseDated(ctx context.Context, rows pgx.Rows) ([]*ItemLinkWithItem, error) {
+	defer rows.Close()
+	links := []*ItemLink{}
+	itemIDs := map[string]bool{}
+	for rows.Next() {
+		er, err := scanEdgeRow(rows)
+		if err != nil {
+			return nil, err
+		}
+		l := linkFromEdge(er)
+		links = append(links, l)
+		itemIDs[l.ItemID] = true
+	}
+	if err := rows.Err(); err != nil {
+		return nil, err
+	}
+	if len(links) == 0 {
+		return nil, nil
+	}
+	// Single bulk fetch for the source-side nodes the rows reference.
+	ids := make([]string, 0, len(itemIDs))
+	for id := range itemIDs {
+		ids = append(ids, id)
+	}
+	gc, err := loadAllProjaxNodes(ctx, r.pool) // simpler than narrow batch
+	if err != nil {
+		return nil, err
+	}
+	out := make([]*ItemLinkWithItem, 0, len(links))
+	for _, l := range links {
+		it := gc.buildItem(l.ItemID)
+		if it == nil {
+			continue
+		}
+		out = append(out, &ItemLinkWithItem{
+			Link:      *l,
+			ItemSlug:  it.Slug,
+			ItemTitle: it.Title,
+			ItemPaths: it.Paths,
+		})
+	}
+	return out, nil
+}
+
+// EncodeDebug serialises a small slice of items for diff-test purposes.
+// Not part of ItemReader; kept here so the parity test in mbrian_test.go
+// can produce deterministic output for the *Store vs MBrianReader diff.
+func EncodeDebug(items []*Item) string {
+	out := make([]map[string]any, 0, len(items))
+	for _, it := range items {
+		out = append(out, map[string]any{
+			"id":         it.ID,
+			"slug":       it.Slug,
+			"title":      it.Title,
+			"paths":      it.Paths,
+			"status":     it.Status,
+			"tags":       it.Tags,
+			"management": it.Management,
+		})
+	}
+	b, _ := json.MarshalIndent(out, "", "  ")
+	return string(b)
+}
--- a/store/mbrian_parity_test.go
+++ b/store/mbrian_parity_test.go
@@ -0,0 +1,251 @@
+package store_test
+
+// Phase 6 Slice B — parity test between the legacy pgx-against-projax-
+// items *Store and the new pgx-against-mbrian *MBrianReader.
+//
+// Skipped without SUPABASE_DATABASE_URL set. When run against the live
+// post-migration database, every comparison should hold: the adapter
+// must be a faithful translator of the migrated graph for projax UI
+// consumers.
+
+import (
+	"context"
+	"errors"
+	"os"
+	"sort"
+	"testing"
+	"time"
+
+	"github.com/jackc/pgx/v5/pgxpool"
+
+	"github.com/m/projax/store"
+)
+
+func newPair(t *testing.T) (*store.Store, *store.MBrianReader, *pgxpool.Pool) {
+	t.Helper()
+	url := os.Getenv("SUPABASE_DATABASE_URL")
+	if url == "" {
+		url = os.Getenv("PROJAX_DB_URL")
+	}
+	if url == "" {
+		t.Skip("set SUPABASE_DATABASE_URL")
+	}
+	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+	defer cancel()
+	pool, err := pgxpool.New(ctx, url)
+	if err != nil {
+		t.Fatalf("pool: %v", err)
+	}
+	if err := pool.Ping(ctx); err != nil {
+		t.Skipf("DB unreachable: %v", err)
+	}
+	return store.New(pool), store.NewMBrianReader(pool), pool
+}
+
+// TestParityListAll: both readers return the same set of items by slug.
+// Field-by-field equality is asserted for slug/title/status/tags/management/
+// public/parent count/paths — the consumer-facing surface.
+func TestParityListAll(t *testing.T) {
+	s, r, pool := newPair(t)
+	defer pool.Close()
+	ctx := context.Background()
+
+	leg, err := s.ListAll(ctx)
+	if err != nil {
+		t.Fatalf("store ListAll: %v", err)
+	}
+	mb, err := r.ListAll(ctx)
+	if err != nil {
+		t.Fatalf("mbrian ListAll: %v", err)
+	}
+	if len(leg) != len(mb) {
+		t.Fatalf("count mismatch: store=%d mbrian=%d", len(leg), len(mb))
+	}
+	// Per the migration brief, two projax-side squatter slugs were
+	// renamed-aside (not deleted) so mBrian could take the canonical
+	// 'work' (area) + 'dania' (project) slugs. Compare every slug that
+	// resolves in BOTH sets; the squatters surface as legacy-only.
+	skip := map[string]bool{"work": true, "dania": true}
+	legBySlug := bySlug(leg)
+	mbBySlug := bySlug(mb)
+	for slug, l := range legBySlug {
+		if skip[slug] {
+			continue
+		}
+		m, ok := mbBySlug[slug]
+		if !ok {
+			t.Errorf("slug %q missing in mBrian set", slug)
+			continue
+		}
+		if l.Title != m.Title {
+			t.Errorf("%s title: store=%q mbrian=%q", slug, l.Title, m.Title)
+		}
+		if l.Status != m.Status {
+			t.Errorf("%s status: store=%q mbrian=%q", slug, l.Status, m.Status)
+		}
+		if !sameSet(l.Tags, m.Tags) {
+			t.Errorf("%s tags: store=%v mbrian=%v", slug, l.Tags, m.Tags)
+		}
+		if !sameSet(l.Management, m.Management) {
+			t.Errorf("%s management: store=%v mbrian=%v", slug, l.Management, m.Management)
+		}
+		if l.Public != m.Public {
+			t.Errorf("%s public: store=%v mbrian=%v", slug, l.Public, m.Public)
+		}
+		if len(l.ParentIDs) != len(m.ParentIDs) {
+			t.Errorf("%s parent count: store=%d mbrian=%d",
+				slug, len(l.ParentIDs), len(m.ParentIDs))
+		}
+		if !sameSet(l.Paths, m.Paths) {
+			t.Errorf("%s paths: store=%v mbrian=%v", slug, l.Paths, m.Paths)
+		}
+	}
+}
+
+// TestParitySpotChecks asserts the 5 spot-check items resolve identically
+// through both readers — root area / single-parent / multi-parent /
+// caldav-linked / public-listing populated.
+func TestParitySpotChecks(t *testing.T) {
+	s, r, pool := newPair(t)
+	defer pool.Close()
+	ctx := context.Background()
+	for _, slug := range []string{"dev", "work", "paliad", "services", "mhome", "fdbck", "dania"} {
+		l, lerr := s.GetByPathOrSlug(ctx, slug)
+		m, merr := r.GetByPathOrSlug(ctx, slug)
+		if lerr != nil && merr != nil {
+			// Both 404 — consistent.
+			continue
+		}
+		if lerr != nil {
+			t.Errorf("%s: store err=%v but mbrian found", slug, lerr)
+			continue
+		}
+		if merr != nil {
+			t.Errorf("%s: mbrian err=%v but store found", slug, merr)
+			continue
+		}
+		if l.Slug != m.Slug || l.Title != m.Title {
+			t.Errorf("%s: shape mismatch store=%+v mbrian=%+v",
+				slug, l.Slug+"/"+l.Title, m.Slug+"/"+m.Title)
+		}
+	}
+}
+
+// TestParityCalDAVLinks: the single caldav-list link must round-trip.
+func TestParityCalDAVLinks(t *testing.T) {
+	s, r, pool := newPair(t)
+	defer pool.Close()
+	ctx := context.Background()
+	leg, err := s.LinksByRefType(ctx, "caldav-list")
+	if err != nil {
+		t.Fatalf("store: %v", err)
+	}
+	mb, err := r.LinksByRefType(ctx, "caldav-list")
+	if err != nil {
+		t.Fatalf("mbrian: %v", err)
+	}
+	if len(leg) != len(mb) {
+		t.Errorf("count: store=%d mbrian=%d", len(leg), len(mb))
+	}
+	// The URLs must round-trip identically. Match by ref_id.
+	legByRef := map[string]*store.ItemLink{}
+	for _, l := range leg {
+		legByRef[l.RefID] = l
+	}
+	for _, m := range mb {
+		l, ok := legByRef[m.RefID]
+		if !ok {
+			t.Errorf("mbrian caldav RefID %q not in store set", m.RefID)
+			continue
+		}
+		if l.Rel != m.Rel {
+			t.Errorf("caldav %s rel: store=%q mbrian=%q", m.RefID, l.Rel, m.Rel)
+		}
+	}
+}
+
+// TestParityGiteaRepoLinks — same parity check for the 37 gitea-repo edges.
+func TestParityGiteaRepoLinks(t *testing.T) {
+	s, r, pool := newPair(t)
+	defer pool.Close()
+	ctx := context.Background()
+	leg, err := s.LinksByRefType(ctx, "gitea-repo")
+	if err != nil {
+		t.Fatalf("store: %v", err)
+	}
+	mb, err := r.LinksByRefType(ctx, "gitea-repo")
+	if err != nil {
+		t.Fatalf("mbrian: %v", err)
+	}
+	if len(leg) != len(mb) {
+		t.Errorf("count: store=%d mbrian=%d", len(leg), len(mb))
+	}
+	legSeen := map[string]bool{}
+	for _, l := range leg {
+		legSeen[l.RefID] = true
+	}
+	for _, m := range mb {
+		if !legSeen[m.RefID] {
+			t.Errorf("mbrian gitea-repo RefID %q not in store set", m.RefID)
+		}
+	}
+}
+
+// TestParityAllTags: tag union must match (modulo ordering).
+func TestParityAllTags(t *testing.T) {
+	s, r, pool := newPair(t)
+	defer pool.Close()
+	ctx := context.Background()
+	leg, err := s.AllTags(ctx)
+	if err != nil {
+		t.Fatalf("store: %v", err)
+	}
+	mb, err := r.AllTags(ctx)
+	if err != nil {
+		t.Fatalf("mbrian: %v", err)
+	}
+	if !sameSet(leg, mb) {
+		t.Errorf("AllTags mismatch:\n  store=%v\n  mbrian=%v", leg, mb)
+	}
+}
+
+// TestParityNotFound: an unknown slug must 404 from both.
+func TestParityNotFound(t *testing.T) {
+	s, r, pool := newPair(t)
+	defer pool.Close()
+	ctx := context.Background()
+	_, le := s.GetByPathOrSlug(ctx, "definitely-not-a-real-slug-zzzzz")
+	_, me := r.GetByPathOrSlug(ctx, "definitely-not-a-real-slug-zzzzz")
+	if !errors.Is(le, store.ErrNotFound) {
+		t.Errorf("store should ErrNotFound, got %v", le)
+	}
+	if !errors.Is(me, store.ErrNotFound) {
+		t.Errorf("mbrian should ErrNotFound, got %v", me)
+	}
+}
+
+// --- helpers ---
+
+func bySlug(items []*store.Item) map[string]*store.Item {
+	out := map[string]*store.Item{}
+	for _, it := range items {
+		out[it.Slug] = it
+	}
+	return out
+}
+
+func sameSet(a, b []string) bool {
+	if len(a) != len(b) {
+		return false
+	}
+	ac := append([]string{}, a...)
+	bc := append([]string{}, b...)
+	sort.Strings(ac)
+	sort.Strings(bc)
+	for i := range ac {
+		if ac[i] != bc[i] {
+			return false
+		}
+	}
+	return true
+}
--- a/web/bulk.go
+++ b/web/bulk.go
@@ -19,7 +19,7 @@ import (
 // matching item, and an action bar that posts to /admin/bulk/apply. The page
 // is intentionally desktop-only — m bulk-edits from a keyboard.
 func (s *Server) handleBulk(w http.ResponseWriter, r *http.Request) {
-	items, err := s.Store.ListAll(r.Context())
+	items, err := s.Items.ListAll(r.Context())
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -29,7 +29,7 @@ func (s *Server) handleBulk(w http.ResponseWriter, r *http.Request) {
 		s.fail(w, r, err)
 		return
 	}
-	allTags, err := s.Store.AllTags(r.Context())
+	allTags, err := s.Items.AllTags(r.Context())
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -377,7 +377,7 @@ func normaliseFormStrings(in []string) []string {
 // the first. Must use the slice form here or the second+ values silently
 // drop on every Apply round-trip.
 func (s *Server) renderBulkList(w http.ResponseWriter, r *http.Request, banner string) {
-	items, err := s.Store.ListAll(r.Context())
+	items, err := s.Items.ListAll(r.Context())
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -387,7 +387,7 @@ func (s *Server) renderBulkList(w http.ResponseWriter, r *http.Request, banner s
 		s.fail(w, r, err)
 		return
 	}
-	allTags, err := s.Store.AllTags(r.Context())
+	allTags, err := s.Items.AllTags(r.Context())
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -445,7 +445,7 @@ func (s *Server) handleBulkChip(w http.ResponseWriter, r *http.Request) {
 		s.fail(w, r, err)
 		return
 	}
-	it, err := s.Store.GetByID(r.Context(), id)
+	it, err := s.Items.GetByID(r.Context(), id)
 	if err != nil {
 		s.fail(w, r, err)
 		return
--- a/web/caldav.go
+++ b/web/caldav.go
@@ -44,11 +44,11 @@ func (s *Server) buildCalDAVOverview(ctx context.Context) (*CalDAVOverview, erro
 	if err != nil {
 		return nil, fmt.Errorf("caldav list: %w", err)
 	}
-	items, err := s.Store.ListAll(ctx)
+	items, err := s.Items.ListAll(ctx)
 	if err != nil {
 		return nil, err
 	}
-	links, err := s.Store.LinksByRefType(ctx, refTypeCalDAV)
+	links, err := s.Items.LinksByRefType(ctx, refTypeCalDAV)
 	if err != nil {
 		return nil, err
 	}
@@ -194,7 +194,7 @@ func (s *Server) handleCalDAVLinkExisting(w http.ResponseWriter, r *http.Request
 		http.Error(w, "caldav not configured", http.StatusServiceUnavailable)
 		return
 	}
-	it, err := s.Store.GetByPath(r.Context(), path)
+	it, err := s.Items.GetByPath(r.Context(), path)
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -245,7 +245,7 @@ func (s *Server) handleCalDAVCreate(w http.ResponseWriter, r *http.Request, path
 		http.Error(w, "caldav not configured", http.StatusServiceUnavailable)
 		return
 	}
-	it, err := s.Store.GetByPath(r.Context(), path)
+	it, err := s.Items.GetByPath(r.Context(), path)
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -306,7 +306,7 @@ func (s *Server) detailTodos(ctx context.Context, item *store.Item) ([]calendarT
 	if s.CalDAV == nil {
 		return nil, nil
 	}
-	links, err := s.Store.LinksByType(ctx, item.ID, refTypeCalDAV)
+	links, err := s.Items.LinksByType(ctx, item.ID, refTypeCalDAV)
 	if err != nil {
 		return nil, err
 	}
@@ -372,7 +372,7 @@ func (s *Server) handleCalDAVTodoAction(w http.ResponseWriter, r *http.Request,
 		http.Error(w, "caldav not configured", http.StatusServiceUnavailable)
 		return
 	}
-	it, err := s.Store.GetByPath(r.Context(), path)
+	it, err := s.Items.GetByPath(r.Context(), path)
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -388,7 +388,7 @@ func (s *Server) handleCalDAVTodoAction(w http.ResponseWriter, r *http.Request,
 	}
 	// Guard: the calendar URL must be linked to this item — otherwise a
 	// crafted form could route writes to arbitrary calendars.
-	links, err := s.Store.LinksByType(r.Context(), it.ID, refTypeCalDAV)
+	links, err := s.Items.LinksByType(r.Context(), it.ID, refTypeCalDAV)
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -541,7 +541,7 @@ func (s *Server) renderTasksSection(w http.ResponseWriter, r *http.Request, it *
 	// here are non-fatal — degrade to an empty picker.
 	var available []caldav.Calendar
 	if s.CalDAV != nil {
-		caldavLinks, lerr := s.Store.LinksByType(r.Context(), it.ID, refTypeCalDAV)
+		caldavLinks, lerr := s.Items.LinksByType(r.Context(), it.ID, refTypeCalDAV)
 		if lerr != nil {
 			s.Logger.Warn("tasks-section caldav links", "path", it.PrimaryPath(), "err", lerr)
 		}
--- a/web/calendar.go
+++ b/web/calendar.go
@@ -209,7 +209,7 @@ func (s *Server) handleCalendar(w http.ResponseWriter, r *http.Request) {
 // lead/trail cells), bins them into per-day cells, and caps each cell at
 // calendarMaxRowsPerCell with the overflow count.
 func (s *Server) buildCalendar(ctx context.Context, q calendarQuery, now time.Time) (*calendarPayload, error) {
-	items, err := s.Store.ListAll(ctx)
+	items, err := s.Items.ListAll(ctx)
 	if err != nil {
 		return nil, err
 	}
--- a/web/dashboard.go
+++ b/web/dashboard.go
@@ -357,7 +357,7 @@ func dashboardTabs(active, filterKey, scope string) []dashboardTab {
 // shapes AND the new per-project rollup so the rollup costs zero extra
 // DAV/Gitea calls.
 func (s *Server) buildDashboard(ctx context.Context, filter TreeFilter) (*dashboardPayload, error) {
-	items, err := s.Store.ListAll(ctx)
+	items, err := s.Items.ListAll(ctx)
 	if err != nil {
 		return nil, err
 	}
@@ -422,7 +422,7 @@ func (s *Server) buildDashboard(ctx context.Context, filter TreeFilter) (*dashbo

 	// --- Recent documents card ---
 	since := now.AddDate(0, 0, -30)
-	docRows, err := s.Store.RecentDocuments(ctx, since, 200)
+	docRows, err := s.Items.RecentDocuments(ctx, since, 200)
 	if err != nil {
 		s.Logger.Warn("dashboard docs", "err", err)
 	}
@@ -486,7 +486,7 @@ func (s *Server) collectStale(ctx context.Context, items []*store.Item, openTask
 		if openTasks[it.ID] > 0 || openIssues[it.ID] > 0 {
 			continue
 		}
-		links, err := s.Store.LinksByType(ctx, it.ID, refTypeGiteaRepo)
+		links, err := s.Items.LinksByType(ctx, it.ID, refTypeGiteaRepo)
 		if err != nil || len(links) == 0 {
 			continue
 		}
@@ -868,7 +868,7 @@ func (s *Server) dashboardTaskWrite(w http.ResponseWriter, r *http.Request, acti
 // pointing at the given URL. Used as the dashboard's write-side ownership
 // guard.
 func (s *Server) calendarLinked(ctx context.Context, calURL string) (bool, error) {
-	links, err := s.Store.LinksByRefType(ctx, refTypeCalDAV)
+	links, err := s.Items.LinksByRefType(ctx, refTypeCalDAV)
 	if err != nil {
 		return false, err
 	}
--- a/web/gitea.go
+++ b/web/gitea.go
@@ -162,7 +162,7 @@ func (s *Server) detailIssues(ctx context.Context, item *store.Item) ([]repoIssu
 	if s.Gitea == nil {
 		return nil, nil
 	}
-	links, err := s.Store.LinksByType(ctx, item.ID, refTypeGiteaRepo)
+	links, err := s.Items.LinksByType(ctx, item.ID, refTypeGiteaRepo)
 	if err != nil {
 		return nil, err
 	}
--- a/web/gitea_writeback.go
+++ b/web/gitea_writeback.go
@@ -22,7 +22,7 @@ func (s *Server) handleIssueAction(w http.ResponseWriter, r *http.Request, path,
 		http.Error(w, "gitea not configured", http.StatusServiceUnavailable)
 		return
 	}
-	it, err := s.Store.GetByPath(r.Context(), path)
+	it, err := s.Items.GetByPath(r.Context(), path)
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -135,7 +135,7 @@ func (s *Server) renderIssuesSection(w http.ResponseWriter, r *http.Request, it
 // to this item via a gitea-repo item_link. Prevents form-crafted writeback
 // against unrelated repos.
 func (s *Server) repoLinkedToItem(ctx context.Context, itemID, repoRef string) bool {
-	links, err := s.Store.LinksByType(ctx, itemID, refTypeGiteaRepo)
+	links, err := s.Items.LinksByType(ctx, itemID, refTypeGiteaRepo)
 	if err != nil {
 		return false
 	}
--- a/web/graph.go
+++ b/web/graph.go
@@ -41,7 +41,7 @@ type graphPayload struct {
 }

 func (s *Server) handleGraph(w http.ResponseWriter, r *http.Request) {
-	items, err := s.Store.ListAll(r.Context())
+	items, err := s.Items.ListAll(r.Context())
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -51,7 +51,7 @@ func (s *Server) handleGraph(w http.ResponseWriter, r *http.Request) {
 		s.fail(w, r, err)
 		return
 	}
-	allTags, err := s.Store.AllTags(r.Context())
+	allTags, err := s.Items.AllTags(r.Context())
 	if err != nil {
 		s.fail(w, r, err)
 		return
--- a/web/links.go
+++ b/web/links.go
@@ -15,7 +15,7 @@ import (
 // note, event_date (YYYY-MM-DD). Anti-forgery isn't a concern at v1 since the
 // trust model is Tailscale-only + cookie auth.
 func (s *Server) handleLinksAdd(w http.ResponseWriter, r *http.Request, path string) {
-	it, err := s.Store.GetByPath(r.Context(), path)
+	it, err := s.Items.GetByPath(r.Context(), path)
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -57,7 +57,7 @@ func (s *Server) handleLinksAdd(w http.ResponseWriter, r *http.Request, path str

 // handleLinksRemove processes POST /i/{path}/links/remove.
 func (s *Server) handleLinksRemove(w http.ResponseWriter, r *http.Request, path string) {
-	it, err := s.Store.GetByPath(r.Context(), path)
+	it, err := s.Items.GetByPath(r.Context(), path)
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -106,7 +106,7 @@ func (s *Server) renderDocumentsSection(w http.ResponseWriter, r *http.Request,
 		http.Redirect(w, r, "/i/"+it.PrimaryPath()+"#documents-section", http.StatusSeeOther)
 		return
 	}
-	docs, err := s.Store.DatedLinks(r.Context(), it.ID)
+	docs, err := s.Items.DatedLinks(r.Context(), it.ID)
 	if err != nil {
 		s.fail(w, r, err)
 		return
--- a/web/server.go
+++ b/web/server.go
@@ -75,7 +75,13 @@ var staticFS embed.FS

 // Server bundles handlers, templates, and the store.
 type Server struct {
-	Store       *store.Store
+	Store *store.Store
+	// Items is the read-path adapter every UI handler / MCP read tool /
+	// aggregator depends on. Phase 6 Slice B introduces it: today the
+	// concrete *Store satisfies the ItemReader interface (legacy path);
+	// after the mBrian backend rollout PROJAX_BACKEND=mbrian wires
+	// *store.MBrianReader here. Writes still flow through Store.
+	Items       store.ItemReader
 	pages       map[string]*template.Template
 	Logger      *slog.Logger
 	Auth        *AuthConfig // nil → no auth (local dev / tests)
@@ -367,7 +373,10 @@ func New(s *store.Store, logger *slog.Logger) (*Server, error) {
 	pages["bulk_chip_mgmt"] = bulkChipMgmt

 	return &Server{
-		Store:       s,
+		Store: s,
+		// Default Items satisfier is *Store itself. main.go can override
+		// post-construction (e.g. PROJAX_BACKEND=mbrian → MBrianReader).
+		Items:       s,
 		pages:       pages,
 		Logger:      logger,
 		dashboard:   cache.NewTTL[*dashboardPayload](dashboardCacheTTL),
@@ -468,12 +477,12 @@ func (s *Server) handleTree(w http.ResponseWriter, r *http.Request) {
 	// Phase 5j slice C: handleTree is reached at /views/tree (system view)
 	// only. The legacy / route 301-redirects via legacyRedirect — see
 	// Routes(). Any 404-on-unknown-path responsibility moved with it.
-	items, err := s.Store.ListAll(r.Context())
+	items, err := s.Items.ListAll(r.Context())
 	if err != nil {
 		s.fail(w, r, err)
 		return
 	}
-	tags, err := s.Store.AllTags(r.Context())
+	tags, err := s.Items.AllTags(r.Context())
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -543,7 +552,7 @@ func (s *Server) handleTree(w http.ResponseWriter, r *http.Request) {
 func (s *Server) linkKindsByItem(ctx context.Context) (map[string]map[string]struct{}, error) {
 	out := map[string]map[string]struct{}{}
 	for _, t := range []string{"caldav-list", "gitea-repo"} {
-		links, err := s.Store.LinksByRefType(ctx, t)
+		links, err := s.Items.LinksByRefType(ctx, t)
 		if err != nil {
 			return nil, err
 		}
@@ -568,11 +577,11 @@ func (s *Server) handleDetail(w http.ResponseWriter, r *http.Request) {
 	// PER URL resolution: try the full path first; if it 404s and the trailing
 	// segment looks like YYMMDD, retry against the shorter path and surface
 	// the date as a render hint to scroll/highlight the matching row.
-	it, err := s.Store.GetByPath(r.Context(), path)
+	it, err := s.Items.GetByPath(r.Context(), path)
 	var highlight *time.Time
 	if errors.Is(err, store.ErrNotFound) {
 		if base, d := parsePER(path); d != nil {
-			if it2, err2 := s.Store.GetByPath(r.Context(), base); err2 == nil {
+			if it2, err2 := s.Items.GetByPath(r.Context(), base); err2 == nil {
 				it, err, highlight = it2, nil, d
 			}
 		}
@@ -596,7 +605,7 @@ func (s *Server) handleDetail(w http.ResponseWriter, r *http.Request) {
 	// are non-fatal — the section falls back to its pre-5j shape.
 	var availableCalendars []caldav.Calendar
 	if s.CalDAV != nil {
-		caldavLinks, lerr := s.Store.LinksByType(r.Context(), it.ID, refTypeCalDAV)
+		caldavLinks, lerr := s.Items.LinksByType(r.Context(), it.ID, refTypeCalDAV)
 		if lerr != nil {
 			s.Logger.Warn("detail caldav links", "path", it.PrimaryPath(), "err", lerr)
 		}
@@ -614,7 +623,7 @@ func (s *Server) handleDetail(w http.ResponseWriter, r *http.Request) {
 	for _, ri := range issues {
 		openTotal += ri.OpenCount
 	}
-	docs, err := s.Store.DatedLinks(r.Context(), it.ID)
+	docs, err := s.Items.DatedLinks(r.Context(), it.ID)
 	if err != nil {
 		s.Logger.Warn("detail docs", "path", it.PrimaryPath(), "err", err)
 	}
@@ -669,7 +678,7 @@ func (s *Server) handleDetailWrite(w http.ResponseWriter, r *http.Request) {
 		s.handleLinksRemove(w, r, base)
 		return
 	}
-	it, err := s.Store.GetByPath(r.Context(), path)
+	it, err := s.Items.GetByPath(r.Context(), path)
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -737,7 +746,7 @@ func (s *Server) handleDetailWrite(w http.ResponseWriter, r *http.Request) {
 // a root mai-managed item under a chosen parent without touching other fields.
 // HTMX-friendly: returns a fragment when HX-Request is set.
 func (s *Server) handleReparent(w http.ResponseWriter, r *http.Request, path string) {
-	it, err := s.Store.GetByPath(r.Context(), path)
+	it, err := s.Items.GetByPath(r.Context(), path)
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -892,7 +901,7 @@ func (s *Server) handleNewForm(w http.ResponseWriter, r *http.Request) {
 	parentPath := r.URL.Query().Get("parent")
 	var parent *store.Item
 	if parentPath != "" {
-		p, err := s.Store.GetByPath(r.Context(), parentPath)
+		p, err := s.Items.GetByPath(r.Context(), parentPath)
 		if err != nil {
 			s.fail(w, r, err)
 			return
@@ -967,7 +976,7 @@ func (s *Server) handleNewSubmit(w http.ResponseWriter, r *http.Request) {
 }

 func (s *Server) handleClassify(w http.ResponseWriter, r *http.Request) {
-	orphans, err := s.Store.MaiOrphans(r.Context())
+	orphans, err := s.Items.MaiOrphans(r.Context())
 	if err != nil {
 		s.fail(w, r, err)
 		return
@@ -993,7 +1002,7 @@ type ParentOption struct {
 }

 func (s *Server) parentOptions(ctx context.Context) ([]ParentOption, error) {
-	items, err := s.Store.ListAll(ctx)
+	items, err := s.Items.ListAll(ctx)
 	if err != nil {
 		return nil, err
 	}
@@ -1065,7 +1074,7 @@ func (s *Server) render(w http.ResponseWriter, r *http.Request, name string, dat
 				}
 			}
 			if needsCount {
-				items, err := s.Store.ListAll(r.Context())
+				items, err := s.Items.ListAll(r.Context())
 				if err == nil {
 					linkKinds, _ := s.linkKindsByItem(r.Context())
 					for _, v := range uv {
--- a/web/timeline.go
+++ b/web/timeline.go
@@ -212,7 +212,7 @@ func (s *Server) handleTimeline(w http.ResponseWriter, r *http.Request) {
 // buildTimeline gathers every dated source, applies the kind/filter narrowing,
 // and groups rows by day in the requested order.
 func (s *Server) buildTimeline(ctx context.Context, q TimelineQuery, now time.Time) (*TimelinePayload, error) {
-	items, err := s.Store.ListAll(ctx)
+	items, err := s.Items.ListAll(ctx)
 	if err != nil {
 		return nil, err
 	}
--- a/web/views.go
+++ b/web/views.go
@@ -100,12 +100,12 @@ func (s *Server) handleViewRender(w http.ResponseWriter, r *http.Request) {
 // dispatch shape. Calendar / timeline view_types fall back to list in
 // slice B; slice D wires their dedicated templates.
 func (s *Server) renderViewPage(w http.ResponseWriter, r *http.Request, v *store.View, filter TreeFilter, viewType, groupBy string) {
-	items, err := s.Store.ListAll(r.Context())
+	items, err := s.Items.ListAll(r.Context())
 	if err != nil {
 		s.fail(w, r, err)
 		return
 	}
-	tags, err := s.Store.AllTags(r.Context())
+	tags, err := s.Items.AllTags(r.Context())
 	if err != nil {
 		s.fail(w, r, err)
 		return
Author	SHA1	Message	Date
mAi	307a898dbd	Merge branch 'mai/kahn/phase-6-sliceB' (Phase 6 Slice B: mBrian-backed read path behind PROJAX_BACKEND switch, parity-green)	2026-05-31 22:22:40 +02:00
mAi	b22f50ca7b	feat(adapter): Phase 6 Slice B — mBrian-backed read path live Per t-projax-6-sliceB-readpath. mBrian migration (m/mBrian#73) is live on msupabase with 65 nodes + 78 child_of + 81 projax-* edges. This commit makes the projax read path source from there behind an env switch. CLIENT ARCH: direct pgxpool against mbrian.* schema (same SUPABASE_DATABASE_URL the projax binary already uses for projax.) — matches flexsiebels/head's cross-coupling pattern. No MCP token plumbing. CONTRACT (all three honoured) - External links are SELF-EDGES (source=target=item, rel='projax-', payload in edges.metadata). linkFromEdge reads the node's outbound projax-* edges; ref_id derived per ref_type from metadata (caldav url, gitea owner/repo, mai-project mai_project_id). - Slugs finalised: 'work'/'dania' resolve to mBrian's canonical nodes; projax-side squatters (renamed-aside, not deleted) are documented in the parity test as legacy-only and skipped from field comparison. - created_at/updated_at NOT preserved — ItemsCreatedInRange orders off metadata.projax.start_time when present, fall back to mBrian created_at. Aggregator surfaces (timeline / dashboard) read off caldav DTSTART + gitea updated_at, so they're unaffected. NEW FILES - store/mbrian.go: MBrianReader concrete impl. Bulk-loads projax- managed nodes + child_of edges in one pair of queries per call, builds a graphContext in memory, derives Paths via ancestor walk (depth-capped at 64 like projax's trigger). Implements every ItemReader method. - store/mbrian_parity_test.go: 5 parity tests against the live db — ListAll field equality (skipping the renamed squatter slugs), spot-check resolves, caldav-list link round-trip, gitea-repo link round-trip, AllTags union, NotFound consistency. All 5 GREEN. - cmd/projax-remap-views/main.go: one-shot tool to rewrite projax.views.filter_json.project_id from old projax uuids to new mBrian uuids using the audit map mBrian dropped (head will relay the path). Dry-run default; --apply commits. Idempotent. - docs/plans/slice-b-views-projectid-gap.md: surfaces the gap + the remediation path. Must run remap BEFORE slice E drops projax.items. CHANGES - store/adapter.go: kept the ItemReader interface + Store assertion; removed the prep stub (replaced by mbrian.go). - web/server.go: Server.Items store.ItemReader field. web.New defaults Items to the concrete Store (legacy path). main.go overrides to MBrianReader when PROJAX_BACKEND=mbrian. - All read-path call sites in web/ swapped from s.Store.<readMethod>( to s.Items.<readMethod>( for the 15 ItemReader methods. MCP tools unchanged (separate scope; can pivot in a follow-up). Writes still flow through s.Store. - cmd/projax/main.go: PROJAX_BACKEND env switch with "store" (default) and "mbrian" values. Logs the choice at startup. Unknown value refuses to start. SMOKE - go build ./... green; go vet green. - go test ./store/ -count=1 — all parity tests pass against live data. - Local server boot with PROJAX_BACKEND=mbrian — backs binding logs "backend=mbrian (read path via store.MBrianReader)" and serves /views/tree (auth wall protects deeper smoke; parity tests cover that surface). PRE-EXISTING failure NOT addressed in this commit: 3 timeline_filter tests in web/ already failed on main (legacy /timeline URL hits the Phase 5j 301 redirect to /views/timeline). No diff vs main in those test files; out of scope for slice B. OUT OF SCOPE FOR SLICE B (deferred): - MCP read tools migration to ItemReader (separate diff, low risk). - Aggregator's LinkLister wired to ItemReader (currently consumes *Store directly through Server.Aggregator()). - views.filter_json.project_id remap RUN — tool ships here, run waits on the head's relay of the audit-map path. - Slice C write-path. Slice D mai-bridge worker. Slice E drop.	2026-05-31 22:20:38 +02:00
mAi	4fdeca8269	Merge branch 'mai/kahn/phase-6-sliceB-prep' (Phase 6: slice-B adapter interface contract + skeleton, no impl)	2026-05-29 15:18:15 +02:00
mAi	9607d4b307	docs+skeleton: Phase 6 Slice B prep — read-path adapter interface contract Per head's parallel-prep brief while m/mBrian#73 (migration script + [schema] node) is being built mBrian-side. NO mBrian-MCP-backed implementation yet — the migration worker may refine the landed node/edge shape and building the impl now risks rework. Built ONLY the parts stable regardless of mBrian internals: 1. CONSUMER INVENTORY (docs/plans/slice-b-adapter-contract.md §1) - Every store.Store read method (15 methods) with signature + semantics - Every call site across web/, internal/aggregate/, mcp/ — table form - Item / ItemLink field-by-field shape contract: which fields come direct from node columns, which from edge-walk, which from metadata-unpack - Direct pgxpool access flagged out-of-scope (admin counts, bulk tx, links event-date update — slice C reworks those) - Views (5j) explicitly NOT in scope per m's Q5=(a) 2. INTERFACE CONTRACT (store/adapter.go) - ItemReader Go interface — 15 methods, pure projax-shaped structs in/out, zero mBrian type leakage - var _ ItemReader = (Store)(nil) compile-time assertion proving the existing pgx-backed Store satisfies the contract today 3. SKELETON (store/adapter.go MBrianReader) - Empty struct (mBrian client choice deferred to slice B impl) - All 15 methods stubbed, return errNotImplementedSliceB - var _ ItemReader = (MBrianReader)(nil) keeps the stubs in lockstep with the interface as slice B grows - Each stub carries a one-line comment naming the §3 gap(s) it resolves at impl time - `go build ./...` green; `go vet ./store/` green 4. GAP FLAGS (docs/plans/slice-b-adapter-contract.md §3) - item_links.rel free-form annotation → mBrian edge.note (add to m/mBrian#73 §1 for the migration script) - ItemLink.RefID per-rel-type extraction rule (caldav URL vs gitea owner/repo vs mai project uuid) - paths[] recomputation cost (per-request memoisation) - AllTags aggregation (full-scan ok at m's scale; tag-graph deferred per m's Q8) - Roots / MaiOrphans "no outbound child_of edge" predicate - ItemsCreatedInRange scoped to projax_origin marker - Item.Source / SourceRefID constant + mai-edge-derived fields - ItemLinkWithItem join shape (two queries + in-memory join vs bulk MCP helper) - Admin counts — recommend adding Counts(ctx) to ItemReader for cohesion Stays parked after this. Slice B IMPL (mBrian-MCP client wiring + per- method bodies + handler rename from s.Store.X to s.Items.X) waits on the migration completing and uuid map landing.	2026-05-29 15:17:24 +02:00